Skip to content
English

Switch theme

Discover▾

Not sure where to start? This week’s hottest software, models and stacks in three minutes.

AI Apps▾

AI software people use: from assistants like ChatGPT to tools for coding, video and workflows, picked by purpose.

For Agents▾

Software for AI: tools that let your agent (Claude Code, Codex, Cursor…) browse, read docs, run code and remember.

Models▾

Choose models, estimate cost, self-host: prices, rankings, benchmarks, free tiers and compute.

Solutions▾

A goal or an error message? Ready-made stacks, steps and fixes by scenario.

Learn▾

A systematic course from zero to the frontier, with an algorithm library, must-read papers and tutorials.

Frontier▾

Daily AI news, weekly highlights, the latest papers and how AI is changing every discipline.

// Models · Self-hosted compute

Open-model VRAM table

Every current open-weight model from the major labs – how much memory it needs at BF16, FP8 and 4-bit, and which GPU or Mac runs it.

Find a model that fits ↓ VRAM calculator →

799 models · 41 families · 1724 official reposdata 2026-10-22official Hugging Face orgs only · VRAM = estimate

01By family

Each family page lists every size with its memory needs.

02Newest releases

By the date the official Hugging Face repository was created.

ModelTypeParams4-bit totalReleased
Ling-3.0-tiny-UBTreeAnt Ling / Ring / MingText LLM183M1.0 GB2026-10-21
DeepSeek-V4.1-Flash-1017DeepSeekVision-language748B / 16B499 GB2026-10-17
NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16NVIDIA NemotronText LLM3.98B3.6 GB2026-10-16
Hy-Memory-30B-A3BTencent Hunyuan / HyText LLM30.1B / 3B21.4 GB2026-10-15
Nemotron-3.5-Rerank-8B-BF16NVIDIA NemotronReranker8.38B6.1 GB2026-10-15
flux-3-image-klein-pupFLUXImage generation9.41B6.8 GB2026-10-14
granite-switch-4.1-8bIBM GraniteText LLM8.78B7.8 GB2026-10-14
granite-switch-4.1-30bIBM GraniteText LLM30B22.7 GB2026-10-14
Ling-3.0-flash-UBTreeAnt Ling / Ring / MingText LLM482M1.5 GB2026-10-14
Ming-3.0-asr-tinyAnt Ling / Ring / MingSpeech-to-text8.87B6.4 GB2026-10-14
flux-3-image-kleinFLUXImage generation5.55B4.2 GB2026-10-13
Ling-3.1-flashAnt Ling / Ring / MingText LLM560B / 25B375 GB2026-10-13

03All models: filter by what fits

Totals include 10% + 0.5GB overhead and the KV cache for one 8K-token request. “Fits” = total ≤ 90% of memory.

ModelTypeParams (total / active)ContextBF16FP84-bitSmallest GPU (4-bit)LicenceReleased
Ling-3.0-tiny-UBTreeAnt Ling / Ring / Ming Text LLM 183M 128K 1.2 GB1.1 GB1.0 GB 8GB GPU Commercial OK 2026-10-21
DeepSeek-V4.1-Flash-1017DeepSeek Vision-language 748B / 16B 1M 1.65 TB875 GB499 GB 8 × H100 (640GB) Commercial OK 2026-10-17
NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16NVIDIA Nemotron Text LLM 3.98B 256K 9.7 GB5.6 GB3.6 GB 8GB GPU Commercial OK 2026-10-16
Hy-Memory-30B-A3BTencent Hunyuan / Hy Text LLM 30.1B / 3B 256K 67.5 GB36.4 GB21.4 GB RTX 4090 24GB Commercial OK 2026-10-15
Nemotron-3.5-Rerank-8B-BF16NVIDIA Nemotron Reranker 8.38B 256K 18.9 GB10.3 GB6.1 GB 8GB GPU Commercial OK 2026-10-15
flux-3-image-klein-pupFLUX Image generation 9.41B 256K 21.2 GB11.5 GB6.8 GB 8GB GPU Commercial OK 2026-10-14
granite-switch-4.1-8bIBM Granite Text LLM 8.78B 128K 21.2 GB12.2 GB7.8 GB 12GB GPU Commercial OK 2026-10-14
granite-switch-4.1-30bIBM Granite Text LLM 30B 128K 68.7 GB37.8 GB22.7 GB RTX 5090 32GB Commercial OK 2026-10-14
Ling-3.0-flash-UBTreeAnt Ling / Ring / Ming Text LLM 482M 128K 2.2 GB1.7 GB1.5 GB 8GB GPU Commercial OK 2026-10-14
Ming-3.0-asr-tinyAnt Ling / Ring / Ming Speech-to-text 8.87B MoE 8K 20.0 GB10.9 GB6.4 GB 8GB GPU Commercial OK 2026-10-14
flux-3-image-kleinFLUX Image generation 5.55B — 12.7 GB7.0 GB4.2 GB 8GB GPU Custom licence – read it 2026-10-13
Ling-3.1-flashAnt Ling / Ring / Ming Text LLM 560B / 25B 256K 1.23 TB656 GB375 GB 8 × H100 (640GB) Commercial OK 2026-10-13
LFM2.5-24B-A2BLiquid LFM Text LLM 24B / 2B 128K 53.4 GB28.7 GB16.7 GB RTX 4090 24GB Custom licence – read it 2026-10-12
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA Nemotron Vision-language 124.5B / 12B — 274 GB146 GB83.5 GB RTX PRO 6000 96GB Commercial OK 2026-10-12
NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16NVIDIA Nemotron Vision-language 74.9B / 9B — 165 GB88.0 GB50.4 GB 2 × RTX 5090 (64GB) Commercial OK 2026-10-09
Qwen-Image-2.1-TurboQwen Image generation 7.12B — 16.2 GB8.8 GB5.2 GB 8GB GPU Non-commercial 2026-10-09
LIDstral-ArabicMistral / Magistral / Devstral Text LLM — — ——— — Commercial OK 2026-10-08
Voxtral-Mini-4B-Realtime-ArabicMistral / Magistral / Devstral Speech-to-text 4.43B 128K 10.2 GB5.7 GB3.5 GB 8GB GPU Licence not stated 2026-10-08
granite-speech-5.0-2bIBM Granite Speech-to-text 2.31B 16K 5.6 GB3.2 GB2.0 GB 8GB GPU Commercial OK 2026-10-07
MiniCPM-V-4.7-35B-A3BMiniCPM / VoxCPM Vision-language 35.2B / 3B 256K 78.1 GB41.8 GB24.2 GB RTX 5090 32GB Commercial OK 2026-10-07
Nemotron-3.5-Embed-1B-BF16NVIDIA Nemotron Embedding 1.55B 256K 3.9 GB2.3 GB1.5 GB 8GB GPU Commercial OK 2026-10-07
Nemotron-3.5-Embed-8B-BF16NVIDIA Nemotron Embedding 8.38B 256K 18.9 GB10.3 GB6.1 GB 8GB GPU Commercial OK 2026-10-07
dots.llm1.baserednote dots Text LLM 142B / 14B 32K 321 GB175 GB104 GB H200 141GB Commercial OK 2026-10-06
MiniCPM-V-4.7-1BMiniCPM / VoxCPM Vision-language 1.55B 256K 4.0 GB2.4 GB1.6 GB 8GB GPU Commercial OK 2026-10-06
ensemble-forecasting-with-ibm-granite-time-seriesIBM Granite Text LLM — — ——— — Licence not stated 2026-10-01
MiMo-V2.6-Flash-MOPDXiaomi MiMo Text LLM 309B / 15B 1M 680 GB362 GB207 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-09-27
MiMo-V2.6-Pro-MOPDXiaomi MiMo Text LLM 1.02T / 42B 1M 2.24 TB1.19 TB681 GB 8 × H200 (1,128GB) Commercial OK 2026-09-27
Ming-3.0-tts-0.5BAnt Ling / Ring / Ming Text-to-speech / audio 1.48B 32K 3.8 GB2.2 GB1.5 GB 8GB GPU Commercial OK 2026-09-22
Ming-3.0-tts-tinyAnt Ling / Ring / Ming Text-to-speech / audio 8.98B / 1.3B 4K 20.3 GB11.0 GB6.5 GB 8GB GPU Commercial OK 2026-09-22
MiMo-V2.6-Distill-Qwen-9BXiaomi MiMo Vision-language 9.41B 256K 21.5 GB11.8 GB7.0 GB 8GB GPU Commercial OK 2026-09-21
MiMo-V2.6-Flash-RLXiaomi MiMo · reasoning Text LLM 309B / 15B 1M 680 GB362 GB207 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-09-21
MiMo-V2.6-Pro-RLXiaomi MiMo · reasoning Text LLM 1.02T / 42B 1M 2.24 TB1.19 TB681 GB 8 × H200 (1,128GB) Commercial OK 2026-09-21
Step-5-PreviewStepFun Step Text LLM 600B / 27B 1M 1.33 TB713 GB412 GB 8 × H100 (640GB) Custom licence – read it 2026-09-20
Ming-Image-0.1-DesignAnt Ling / Ring / Ming Image generation 6.16B MoE 32K 14.0 GB7.7 GB4.6 GB 8GB GPU Commercial OK 2026-09-17
embeddinggemma-2Gemma Embedding 744M 256K 2.1 GB1.4 GB1.0 GB 8GB GPU Commercial OK 2026-09-14
Qwen-Image-2.1Qwen Image generation 7.12B — 16.2 GB8.8 GB5.2 GB 8GB GPU Non-commercial 2026-09-14
Intern-S2-397BInternLM / Intern-S Vision-language 403.4B MoE 256K 888 GB472 GB270 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-09-13
Atria-Dawn-PreviewInternLM / Intern-S Text LLM 753.3B MoE 1M 1.66 TB882 GB504 GB 8 × H100 (640GB) Commercial OK 2026-09-11
DeepSeek-V4.1-FlashDeepSeek Vision-language 748B / 16B 1M 1.65 TB875 GB499 GB 8 × H100 (640GB) Commercial OK 2026-09-10
MiniCPM5-2BMiniCPM / VoxCPM Text LLM 2.52B 128K 6.4 GB3.8 GB2.5 GB 8GB GPU Commercial OK 2026-09-06
Ling-3.0-flash-VLAnt Ling / Ring / Ming Vision-language 124B / 5.5B 128K 274 GB146 GB83.6 GB RTX PRO 6000 96GB Commercial OK 2026-09-04
Ming-3.0-audio-flashAnt Ling / Ring / Ming Text-to-speech / audio 126B / 6.5B 128K 278 GB148 GB84.5 GB RTX PRO 6000 96GB Commercial OK 2026-09-03
Nemotron-3.5-ParseNVIDIA Nemotron Vision-language 631M 16K 2.1 GB1.4 GB1.1 GB 8GB GPU Commercial OK 2026-09-03
VibeVoice-ASR-Streaming-1.5BMicrosoft VibeVoice / Fara / Harrier Speech-to-text 2.81B — 6.7 GB3.8 GB2.4 GB 8GB GPU Commercial OK 2026-09-02
VibeVoice-ASR-Streaming-7BMicrosoft VibeVoice / Fara / Harrier Speech-to-text 8.67B — 19.6 GB10.6 GB6.3 GB 8GB GPU Commercial OK 2026-09-02
Nemotron-3-Labs-Ultra-Math-RLNVIDIA Nemotron · reasoning Text LLM 550B / 55B 256K 1.21 TB643 GB367 GB 8 × H100 (640GB) Commercial OK 2026-09-02
Hy4-previewTencent Hunyuan / Hy Text LLM 770B / 49B 1M 1.70 TB901 GB515 GB 8 × H100 (640GB) Commercial OK 2026-08-27
GLM-5.3GLM / CogVideoX (Z.ai) Text LLM 753.3B MoE 1M 1.66 TB882 GB504 GB 8 × H100 (640GB) Custom licence – read it 2026-08-25
GLM-5.3-FlashGLM / CogVideoX (Z.ai) Vision-language 320B / 18B 1M 705 GB375 GB214 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-08-25
Qwen3.8-Flash-NextQwen Vision-language 180B / 6B 256K 397 GB211 GB121 GB H200 141GB Custom licence – read it 2026-08-24
granite-speech-5.0-470m-turboctc-ncIBM Granite Speech-to-text 473M — 1.5 GB1.1 GB0.8 GB 8GB GPU Non-commercial 2026-08-17
DeepSeek-V4-Pro-0813DeepSeek Text LLM 1.65T MoE 1M 3.63 TB1.93 TB1.10 TB 8 × B200 (1,440GB) Commercial OK 2026-08-13
LFM2.5-VL-3BLiquid LFM Vision-language 3.12B 32K 7.5 GB4.3 GB2.7 GB 8GB GPU Custom licence – read it 2026-08-11
North-Micro-Vision-InstructCohere Command / Aya Vision-language 2.48B 488K 6.6 GB4.0 GB2.7 GB 8GB GPU Commercial OK 2026-08-10
Ling-3.0-tinyAnt Ling / Ring / Ming Text LLM 7.9B / 1.3B 128K 18.1 GB10.0 GB6.0 GB 8GB GPU Commercial OK 2026-08-10
Qwen3.8-2.4T-A95BQwen Text LLM 2.4T / 95B 256K 5.28 TB2.81 TB1.60 TB 8 × B300 (2,304GB) Custom licence – read it 2026-08-08
granite-4.2-3bIBM Granite Text LLM 3.66B 128K 9.2 GB5.4 GB3.6 GB 8GB GPU Commercial OK 2026-08-07
granite-4.2-8bIBM Granite Text LLM 8.79B 128K 21.2 GB12.1 GB7.7 GB 12GB GPU Commercial OK 2026-08-07
granite-4.2-30bIBM Granite Text LLM 29.3B 128K 67.1 GB36.9 GB22.2 GB RTX 5090 32GB Commercial OK 2026-08-07
MiniMax-Music3MiniMax Text-to-speech / audio 2.43B — 5.9 GB3.3 GB2.1 GB 8GB GPU Licence not stated 2026-08-07
Wan2.2-Animate-2-14B-DiffusersWan Video generation 16.4B — 36.6 GB19.7 GB11.4 GB 16GB GPU Commercial OK 2026-08-06
Wan2.2-Animate-2-14B-Distilled-DiffusersWan Video generation 16.4B — 36.6 GB19.7 GB11.4 GB 16GB GPU Commercial OK 2026-08-06
Qwen3.8-27BQwen Vision-language 27.8B 256K 62.2 GB33.5 GB19.6 GB RTX 4090 24GB Commercial OK 2026-08-05
granite-speech-5.0-470m-turboctcIBM Granite Speech-to-text 473M — 1.5 GB1.1 GB0.8 GB 8GB GPU Commercial OK 2026-08-04
Ling-3.0-flashAnt Ling / Ring / Ming Text LLM 124B / 5.1B 256K 274 GB146 GB83.6 GB RTX PRO 6000 96GB Commercial OK 2026-08-02
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA Nemotron Text LLM 31.6B / 3B 256K 70.0 GB37.5 GB21.6 GB RTX 4090 24GB Commercial OK 2026-08-01
DeepSeek-V4-Flash-0731DeepSeek Text LLM 304.2B MoE 1M 670 GB356 GB203 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-07-31
LongCat-Flash-Lite-SparseMeituan LongCat Text LLM 69B / 3B 960K 152 GB81.3 GB46.6 GB 2 × RTX 5090 (64GB) Commercial OK 2026-07-31
Intern-S2-MobiusInternLM / Intern-S Vision-language 36B MoE 256K 79.8 GB42.7 GB24.7 GB RTX 5090 32GB Commercial OK 2026-07-29
NVIDIA-NemotronLabs-VoiceChat-11BNVIDIA Nemotron Text LLM 11.1B — 24.9 GB13.5 GB7.9 GB 12GB GPU Commercial OK 2026-07-29
LFM2.5-2.6BLiquid LFM Text LLM 2.7B 128K 6.6 GB3.8 GB2.4 GB 8GB GPU Custom licence – read it 2026-07-28
MiniMax-H3MiniMax Video generation 33.1B — 73.4 GB39.2 GB22.6 GB RTX 5090 32GB Custom licence – read it 2026-07-28
Inkling-SmallThinking Machines Inkling Vision-language 276B / 12B — 608 GB323 GB185 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-07-27
LTX-2.5LTX Video Video generation — — ——— — Custom licence – read it 2026-07-23
Fara1.5-4BMicrosoft VibeVoice / Fara / Harrier Vision-language 4.54B 256K 10.8 GB6.1 GB3.8 GB 8GB GPU Commercial OK 2026-07-17
Fara1.5-27BMicrosoft VibeVoice / Fara / Harrier Vision-language 27.4B 256K 61.2 GB33.0 GB19.3 GB RTX 4090 24GB Commercial OK 2026-07-17
Shieldstral-1.0-3BMistral / Magistral / Devstral Safety / guard 3.85B 256K 9.8 GB5.9 GB3.9 GB 8GB GPU Commercial OK 2026-07-16
InklingThinking Machines Inkling Vision-language 975B / 41B — 2.15 TB1.14 TB651 GB 8 × H200 (1,128GB) Commercial OK 2026-07-14
jina-reranker-v3.5Jina embeddings & rerankers Reranker 597M 128K 1.8 GB1.2 GB0.9 GB 8GB GPU Non-commercial 2026-07-14
Nemotron-3-Embed-1B-BF16NVIDIA Nemotron Embedding 1.14B 256K 3.0 GB1.8 GB1.3 GB 8GB GPU Commercial OK 2026-07-14
Nemotron-3-Embed-8B-BF16NVIDIA Nemotron Embedding 7.95B 256K 18.0 GB9.8 GB5.8 GB 8GB GPU Commercial OK 2026-07-14
Laguna-S-2.1poolside Laguna · coding Text LLM 118B / 8B 1M 261 GB139 GB79.7 GB RTX PRO 6000 96GB Commercial OK 2026-07-13
LFM2.5-Encoder-350M-DiffusionLiquid LFM · coding Text LLM 354M 125K 1.4 GB1.0 GB0.8 GB 8GB GPU Custom licence – read it 2026-07-10
Wan-Dancer-14BWan Video generation 14B — 31.3 GB16.9 GB9.8 GB 12GB GPU Commercial OK 2026-07-10
Nemotron-Labs-Audex-2BNVIDIA Nemotron Text LLM 2B — 4.9 GB2.8 GB1.8 GB 8GB GPU Non-commercial 2026-07-06
LongCat-2.0Meituan LongCat Text LLM 1.6T / 48B 256K 3.52 TB1.87 TB1.07 TB 8 × B200 (1,440GB) Commercial OK 2026-07-05
Hy3Tencent Hunyuan / Hy Text LLM 295B / 21B 256K 652 GB348 GB200 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-07-02
NVIDIA-Nemotron-Parse-2.0NVIDIA Nemotron Vision-language 903M — 2.5 GB1.6 GB1.1 GB 8GB GPU Commercial OK 2026-06-30
LFM2.5-230MLiquid LFM Text LLM 230M 125K 1.1 GB0.9 GB0.8 GB 8GB GPU Custom licence – read it 2026-06-24
Laguna-XS-2.1poolside Laguna · coding Text LLM 33B / 3B 256K 73.5 GB39.5 GB22.9 GB RTX 5090 32GB Commercial OK 2026-06-20
GLM-5.2GLM / CogVideoX (Z.ai) Text LLM 753.3B MoE 1M 1.66 TB882 GB504 GB 8 × H100 (640GB) Commercial OK 2026-06-16
Laguna-M.1poolside Laguna · coding Text LLM 225B / 23B 256K 498 GB266 GB153 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-06-15
Kimi-K3Kimi Vision-language 2.8T / 104B 1M 6.16 TB3.27 TB1.87 TB 8 × B300 (2,304GB) Custom licence – read it 2026-06-13
Kimi-K2.7-CodeKimi · coding Vision-language 1T / 32B 256K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial with conditions 2026-06-11
diffusiongemma-26B-A4B-itGemma Vision-language 25.8B / 4B 256K 57.9 GB31.2 GB18.3 GB RTX 4090 24GB Commercial OK 2026-06-09
SCAIL-2GLM / CogVideoX (Z.ai) Video generation — — ——— — Commercial OK 2026-06-09
North-Mini-Code-1.0Cohere Command / Aya · coding Text LLM 30B / 3B 488K 67.0 GB36.1 GB21.0 GB RTX 4090 24GB Commercial OK 2026-06-05
gemma-4-12B-it-qat-q4_0-unquantizedGemma Omni / audio LLM 12B 256K 27.7 GB15.3 GB9.3 GB 12GB GPU Commercial OK 2026-06-04
NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA Nemotron Text LLM 560.5B / 55B 256K 1.23 TB656 GB374 GB 8 × H100 (640GB) Commercial OK 2026-06-03
MiniMax-M3MiniMax Vision-language 428B / 23B 1M 942 GB501 GB286 GB 4 × RTX PRO 6000 (384GB) Custom licence – read it 2026-06-02
LFM2.5-8B-A1BLiquid LFM Text LLM 8.47B / 1B 125K 19.2 GB10.5 GB6.2 GB 8GB GPU Custom licence – read it 2026-05-28
LFM2.5-1.2B-JP-202606Liquid LFM Text LLM 1.17B 125K 3.2 GB2.0 GB1.4 GB 8GB GPU Custom licence – read it 2026-05-26
Fun-ASR-Nano-2512-hfFunAudioLLM speech Speech-to-text 830M 40K 2.3 GB1.5 GB1.1 GB 8GB GPU Commercial OK 2026-05-24
gemma-4-12B-itGemma Omni / audio LLM 12B 256K 27.7 GB15.3 GB9.3 GB 12GB GPU Commercial OK 2026-05-23
Step-3.7-FlashStepFun Step Vision-language 198B / 11B 256K 440 GB236 GB136 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-05-23
MiniCPM5-1BMiniCPM / VoxCPM Text LLM 1.08B 128K 3.1 GB2.0 GB1.4 GB 8GB GPU Commercial OK 2026-05-21
LFM2.5-ColBERT-350MLiquid LFM Embedding 353M 125K 1.3 GB0.9 GB0.7 GB 8GB GPU Custom licence – read it 2026-05-20
stable-audio-3-mediumStable Diffusion / Stable Audio Text-to-speech / audio 2.31B — 5.6 GB3.2 GB2.0 GB 8GB GPU Commercial with conditions 2026-05-17
stable-audio-3-small-musicStable Diffusion / Stable Audio Text-to-speech / audio 568M — 1.7 GB1.2 GB0.9 GB 8GB GPU Commercial with conditions 2026-05-17
stable-audio-3-small-sfxStable Diffusion / Stable Audio Text-to-speech / audio 568M — 1.7 GB1.2 GB0.9 GB 8GB GPU Commercial with conditions 2026-05-17
Intern-S2-PreviewInternLM / Intern-S Vision-language 36.1B MoE 256K 80.1 GB42.9 GB24.7 GB RTX 5090 32GB Commercial OK 2026-05-15
Ring-2.6-1TAnt Ling / Ring / Ming · reasoning Text LLM 1.03T MoE 128K 2.26 TB1.20 TB685 GB 8 × H200 (1,128GB) Commercial OK 2026-05-14
Fara1.5-9BMicrosoft VibeVoice / Fara / Harrier Vision-language 9.41B 256K 21.5 GB11.8 GB7.0 GB 8GB GPU Commercial OK 2026-05-12
command-a-plus-05-2026-bf16Cohere Command / Aya Vision-language 218B / 25B 195K 481 GB256 GB147 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-05-11
Hy-MT2-1.8BTencent Hunyuan / Hy Translation 2.04B 256K 5.5 GB3.4 GB2.4 GB 8GB GPU Commercial OK 2026-05-11
Hy-MT2-7BTencent Hunyuan / Hy Translation 8.03B 256K 19.2 GB11.0 GB6.9 GB 8GB GPU Commercial OK 2026-05-11
Hy-MT2-30B-A3BTencent Hunyuan / Hy Translation 30.1B / 3B 256K 67.5 GB36.4 GB21.4 GB RTX 4090 24GB Commercial OK 2026-05-11
MiniCPM-V-4.6-ThinkingMiniCPM / VoxCPM · reasoning Vision-language 1.3B 256K 3.5 GB2.1 GB1.5 GB 8GB GPU Commercial OK 2026-05-08
Nemotron-Labs-Diffusion-VLM-8BNVIDIA Nemotron Vision-language 8.92B 256K 21.3 GB12.1 GB7.6 GB 12GB GPU Non-commercial 2026-05-08
LFM2.5-Embedding-350MLiquid LFM Embedding 354M 125K 1.3 GB0.9 GB0.7 GB 8GB GPU Custom licence – read it 2026-05-05
Molmo2-ERAi2 OLMo / Molmo Vision-language 4.85B 16K 12.4 GB7.4 GB4.9 GB 8GB GPU Commercial OK 2026-05-04
gemma-4-E4B-it-qat-q4_0-unquantizedGemma Omni / audio LLM 7.94B 128K 18.1 GB9.9 GB5.9 GB 8GB GPU Commercial OK 2026-04-30
gemma-4-E2B-it-qat-q4_0-unquantizedGemma Omni / audio LLM 5.1B 128K 11.8 GB6.5 GB4.0 GB 8GB GPU Commercial OK 2026-04-29
gemma-4-26B-A4B-it-qat-q4_0-unquantizedGemma Vision-language 25.8B / 4B 256K 57.8 GB31.2 GB18.3 GB RTX 4090 24GB Commercial OK 2026-04-29
Ling-2.6-1TAnt Ling / Ring / Ming Text LLM 1.03T MoE 256K 2.26 TB1.20 TB685 GB 8 × H200 (1,128GB) Commercial OK 2026-04-29
gemma-4-31B-it-qat-q4_0-unquantizedGemma Vision-language 31.3B 256K 71.5 GB39.2 GB23.5 GB RTX 5090 32GB Commercial OK 2026-04-28
Ling-2.6-flashAnt Ling / Ring / Ming Text LLM 104B / 7.4B 128K 230 GB122 GB70.2 GB H100 80GB Commercial OK 2026-04-28
MiMo-V2.5Xiaomi MiMo Text LLM 310B / 15B 1M 683 GB363 GB207 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-04-27
MiMo-V2.5-ProXiaomi MiMo Text LLM 1.02T / 42B 1M 2.24 TB1.19 TB681 GB 8 × H200 (1,128GB) Commercial OK 2026-04-27
Laguna-XS.2poolside Laguna · coding Text LLM 33B / 3B 256K 73.5 GB39.5 GB22.9 GB RTX 5090 32GB Commercial OK 2026-04-23
MiMo-V2.5-ASRXiaomi MiMo Speech-to-text 7.62B 8K 17.3 GB9.4 GB5.6 GB 8GB GPU Commercial OK 2026-04-23
DeepSeek-V4-FlashDeepSeek Text LLM 284B / 13B 1M 625 GB332 GB190 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-04-22
DeepSeek-V4-ProDeepSeek Text LLM 1.6T / 49B 1M 3.52 TB1.87 TB1.07 TB 8 × B200 (1,440GB) Commercial OK 2026-04-22
Nemotron-Labs-Diffusion-14BNVIDIA Nemotron Text LLM 13.5B 256K 31.6 GB17.6 GB10.8 GB 12GB GPU Commercial with conditions 2026-04-22
Qwen3.6-27BQwen Vision-language 27.8B 256K 62.2 GB33.5 GB19.6 GB RTX 4090 24GB Commercial OK 2026-04-21
granite-embedding-97m-multilingual-r2IBM Granite Embedding 97M 32K 0.7 GB0.6 GB0.6 GB 8GB GPU Commercial OK 2026-04-20
granite-embedding-311m-multilingual-r2IBM Granite Embedding 312M 32K 1.2 GB0.9 GB0.7 GB 8GB GPU Commercial OK 2026-04-20
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16NVIDIA Nemotron · reasoning Omni / audio LLM 33B / 3B 256K 73.2 GB39.1 GB22.6 GB RTX 5090 32GB Commercial with conditions 2026-04-20
granite-guardian-4.1-8bIBM Granite Safety / guard 8.38B 128K 20.3 GB11.6 GB7.4 GB 12GB GPU Commercial OK 2026-04-16
granite-speech-4.1-2bIBM Granite Speech-to-text 2.31B 4K 5.6 GB3.2 GB2.0 GB 8GB GPU Commercial OK 2026-04-16
granite-speech-4.1-2b-plusIBM Granite Speech-to-text 2.11B 4K 5.1 GB3.0 GB1.9 GB 8GB GPU Commercial OK 2026-04-16
granite-vision-4.1-4bIBM Granite Vision-language 4B 128K 10.0 GB5.8 GB3.8 GB 8GB GPU Commercial OK 2026-04-16
Qwen3.6-35B-A3BQwen Vision-language 36B / 3B 256K 79.8 GB42.7 GB24.6 GB RTX 5090 32GB Commercial OK 2026-04-15
Kimi-K2.6Kimi Vision-language 1T / 32B 256K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial with conditions 2026-04-14
Hy3-previewTencent Hunyuan / Hy Text LLM 295B / 21B 256K 652 GB348 GB200 GB 4 × RTX PRO 6000 (384GB) Custom licence – read it 2026-04-13
MiniCPM-V-4.6MiniCPM / VoxCPM Vision-language 1.3B 256K 3.5 GB2.1 GB1.5 GB 8GB GPU Commercial OK 2026-04-13
MiniMax-M2.7MiniMax Text LLM 228.7B MoE 200K 506 GB270 GB155 GB 4 × RTX PRO 6000 (384GB) Custom licence – read it 2026-04-09
LFM2.5-VL-450MLiquid LFM Vision-language 449M 125K 1.6 GB1.1 GB0.9 GB 8GB GPU Custom licence – read it 2026-04-08
ERNIE-ImageBaidu ERNIE Image generation 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Commercial OK 2026-04-07
parakeet-unified-en-0.6bNVIDIA Parakeet / Canary speech Speech-to-text 600M — 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial with conditions 2026-04-07
granite-4.1-3bIBM Granite Text LLM 3.4B 128K 8.7 GB5.1 GB3.4 GB 8GB GPU Commercial OK 2026-04-06
granite-4.1-8bIBM Granite Text LLM 8.79B 128K 21.2 GB12.1 GB7.7 GB 12GB GPU Commercial OK 2026-04-06
granite-4.1-30bIBM Granite Text LLM 28.9B 128K 66.2 GB36.4 GB21.9 GB RTX 5090 32GB Commercial OK 2026-04-06
GLM-5.1GLM / CogVideoX (Z.ai) Text LLM 753.9B MoE 198K 1.66 TB882 GB504 GB 8 × H100 (640GB) Commercial OK 2026-04-03
VoxCPM2MiniCPM / VoxCPM Text-to-speech / audio 2.29B — 5.5 GB3.2 GB2.0 GB 8GB GPU Commercial OK 2026-04-03
ERNIE-Image-TurboBaidu ERNIE Image generation 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Commercial OK 2026-04-02
jina-embeddings-v5-omni-nanoJina embeddings & rerankers Embedding 986M 8K 2.7 GB1.7 GB1.2 GB 8GB GPU Non-commercial 2026-04-01
jina-embeddings-v5-omni-smallJina embeddings & rerankers Embedding 1.63B 32K 4.1 GB2.4 GB1.6 GB 8GB GPU Non-commercial 2026-04-01
nemotron-ocr-v2NVIDIA Nemotron Vision-language — — ——— — Commercial with conditions 2026-04-01
jina-embeddings-v5-omni-nano-retrievalJina embeddings & rerankers Embedding 948M 8K 2.6 GB1.6 GB1.1 GB 8GB GPU Non-commercial 2026-03-31
jina-embeddings-v5-omni-small-retrievalJina embeddings & rerankers Embedding 1.56B 32K 3.9 GB2.3 GB1.5 GB 8GB GPU Non-commercial 2026-03-31
LFM2.5-350MLiquid LFM Text LLM 354M 125K 1.4 GB1.0 GB0.8 GB 8GB GPU Custom licence – read it 2026-03-31
Mistral-Medium-3.5-128BMistral / Magistral / Devstral Text LLM 127.7B 256K 284 GB153 GB88.6 GB H200 141GB Custom licence – read it 2026-03-31
LongCat-AudioDiT-1BMeituan LongCat Text-to-speech / audio 1.42B — 3.6 GB2.2 GB1.4 GB 8GB GPU Commercial OK 2026-03-30
LongCat-AudioDiT-3.5BMeituan LongCat Text-to-speech / audio 3.83B — 8.9 GB5.0 GB3.1 GB 8GB GPU Commercial OK 2026-03-30
harrier-oss-v1-270mMicrosoft VibeVoice / Fara / Harrier Embedding 268M 32K 1.1 GB0.8 GB0.7 GB 8GB GPU Commercial OK 2026-03-30
harrier-oss-v1-0.6bMicrosoft VibeVoice / Fara / Harrier Embedding 596M 32K 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2026-03-30
harrier-oss-v1-27bMicrosoft VibeVoice / Fara / Harrier Embedding 27B 128K 59.9 GB32.1 GB18.5 GB RTX 4090 24GB Commercial OK 2026-03-30
MolmoWeb-4BAi2 OLMo / Molmo Vision-language 4.85B 10K 12.4 GB7.4 GB4.9 GB 8GB GPU Commercial OK 2026-03-20
Nemotron-Cascade-2-30B-A3BNVIDIA Nemotron Text LLM 31.6B / 3B 256K 70.0 GB37.5 GB21.6 GB RTX 4090 24GB Commercial with conditions 2026-03-18
Nemotron-Labs-Diffusion-8BNVIDIA Nemotron Text LLM 8.49B 256K 20.3 GB11.6 GB7.3 GB 12GB GPU Commercial with conditions 2026-03-18
MolmoPoint-8BAi2 OLMo / Molmo Vision-language 8.68B 37K 20.8 GB11.9 GB7.5 GB 12GB GPU Commercial OK 2026-03-16
gemma-4-26B-A4B-itGemma Vision-language 25.8B / 4B 256K 57.8 GB31.2 GB18.3 GB RTX 4090 24GB Commercial OK 2026-03-11
gemma-4-31B-itGemma Vision-language 31.3B 256K 71.5 GB39.2 GB23.5 GB RTX 5090 32GB Commercial OK 2026-03-11
granite-speech-4.1-2b-narIBM Granite Embedding 2.25B 4K 5.5 GB3.1 GB2.0 GB 8GB GPU Commercial OK 2026-03-10
NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA Nemotron Text LLM 123.6B / 12B 256K 273 GB145 GB83.0 GB RTX PRO 6000 96GB Commercial with conditions 2026-03-10
FLUX.2-klein-9b-kvFLUX Image generation 9.08B — 20.5 GB11.1 GB6.6 GB 8GB GPU Non-commercial 2026-03-09
NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA Nemotron Text LLM 3.97B 256K 9.4 GB5.3 GB3.3 GB 8GB GPU Commercial with conditions 2026-03-07
LTX-2.3LTX Video Video generation — — ——— — Custom licence – read it 2026-03-04
granite-4.0-3b-visionIBM Granite Vision-language 4B 128K 10.0 GB5.8 GB3.8 GB 8GB GPU Commercial OK 2026-03-03
gemma-4-E2B-itGemma Omni / audio LLM 5.12B 128K 11.8 GB6.6 GB4.0 GB 8GB GPU Commercial OK 2026-03-02
gemma-4-E4B-itGemma Omni / audio LLM 8B 128K 18.2 GB10.0 GB6.0 GB 8GB GPU Commercial OK 2026-03-02
Nemotron-Labs-Diffusion-3BNVIDIA Nemotron Text LLM 3.83B 256K 9.8 GB5.8 GB3.9 GB 8GB GPU Commercial with conditions 2026-03-02
Qwen3.5-0.8BQwen Vision-language 873M 256K 2.5 GB1.6 GB1.2 GB 8GB GPU Commercial OK 2026-02-28
Qwen3.5-2BQwen Vision-language 2.27B 256K 5.6 GB3.3 GB2.1 GB 8GB GPU Commercial OK 2026-02-28
granite-4.0-1b-speechIBM Granite Speech-to-text 2.31B 4K 5.6 GB3.2 GB2.0 GB 8GB GPU Commercial OK 2026-02-27
Qwen3.5-4BQwen Vision-language 4.66B 256K 11.0 GB6.2 GB3.9 GB 8GB GPU Commercial OK 2026-02-27
Qwen3.5-9BQwen Vision-language 9.65B 256K 22.0 GB12.1 GB7.2 GB 8GB GPU Commercial OK 2026-02-27
LFM2-24B-A2BLiquid LFM Text LLM 23.8B / 2B 125K 53.1 GB28.5 GB16.6 GB RTX 4090 24GB Custom licence – read it 2026-02-24
Qwen3.5-27BQwen Vision-language 27.8B 256K 62.2 GB33.5 GB19.6 GB RTX 4090 24GB Commercial OK 2026-02-24
Qwen3.5-35B-A3BQwen Vision-language 36B / 3B 256K 79.8 GB42.7 GB24.6 GB RTX 5090 32GB Commercial OK 2026-02-24
Qwen3.5-122B-A10BQwen Vision-language 125.1B / 10B 256K 276 GB147 GB84.1 GB RTX PRO 6000 96GB Commercial OK 2026-02-24
NVIDIA-Nemotron-Parse-v1.2NVIDIA Nemotron Vision-language 936M — 2.6 GB1.6 GB1.1 GB 8GB GPU Commercial with conditions 2026-02-18
Qwen3.5-397B-A17BQwen Vision-language 403.4B / 17B 256K 888 GB472 GB270 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-02-16
Ling-2.5-1TAnt Ling / Ring / Ming Text LLM 1T / 63B 128K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial OK 2026-02-14
tiny-aya-baseCohere Command / Aya Text LLM 3.35B — 7.9 GB4.4 GB2.7 GB 8GB GPU Non-commercial 2026-02-13
tiny-aya-globalCohere Command / Aya Text LLM 3.35B — 7.9 GB4.4 GB2.7 GB 8GB GPU Non-commercial 2026-02-13
MiniMax-M2.5MiniMax Text LLM 228.7B MoE 192K 506 GB270 GB155 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2026-02-12
GLM-5GLM / CogVideoX (Z.ai) Text LLM 744B / 40B 198K 1.64 TB871 GB497 GB 8 × H100 (640GB) Commercial OK 2026-02-11
Ming-omni-tts-0.5BAnt Ling / Ring / Ming Text-to-speech / audio 1.62B 32K 4.1 GB2.4 GB1.6 GB 8GB GPU Commercial OK 2026-02-11
MiniCPM-SALAMiniCPM / VoxCPM Text LLM 9.48B 512K 21.6 GB11.8 GB7.1 GB 8GB GPU Commercial OK 2026-02-11
Ming-flash-omni-2.0Ant Ling / Ring / Ming Omni / audio LLM 100B / 6B 32K 221 GB118 GB67.7 GB H100 80GB Commercial OK 2026-02-10
Ring-2.5-1TAnt Ling / Ring / Ming · reasoning Text LLM 1.01T MoE 128K 2.23 TB1.18 TB676 GB 8 × H200 (1,128GB) Commercial OK 2026-02-10
NVIDIA-Nemotron-Nano-9B-v2-JapaneseNVIDIA Nemotron Text LLM 8.89B 128K 20.2 GB11.0 GB6.6 GB 8GB GPU Commercial with conditions 2026-02-04
LongCat-Image-Edit-TurboMeituan LongCat Image generation — — ——— — Commercial OK 2026-02-03
MiniCPM-o-4_5MiniCPM / VoxCPM Omni / audio LLM 9.37B 40K 22.3 GB12.7 GB8.0 GB 12GB GPU Commercial OK 2026-02-03
granitelib-guardian-r1.0IBM Granite · reasoning Safety / guard — — ——— — Commercial OK 2026-02-02
Intern-S1-ProInternLM / Intern-S Vision-language 1T / 22B 256K 2.20 TB1.17 TB669 GB 8 × H200 (1,128GB) Commercial OK 2026-02-02
Step-3.5-FlashStepFun Step Text LLM 196B / 11B 256K 436 GB233 GB135 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-02-01
GLM-OCRGLM / CogVideoX (Z.ai) Vision-language 1.32B 128K 4.0 GB2.6 GB1.9 GB 8GB GPU Commercial OK 2026-01-30
Qwen3-Coder-NextQwen · coding Text LLM 80B / 3B 256K 177 GB94.2 GB54.1 GB 2 × RTX 5090 (64GB) Commercial OK 2026-01-30
Olmo-Hybrid-7BAi2 OLMo / Molmo Text LLM 7.43B 64K 17.9 GB10.2 GB6.5 GB 8GB GPU Commercial OK 2026-01-28
Qwen3-ASR-0.6BQwen Speech-to-text 938M 64K 2.6 GB1.6 GB1.1 GB 8GB GPU Commercial OK 2026-01-28
Qwen3-ASR-1.7BQwen Speech-to-text 2.35B 64K 5.7 GB3.2 GB2.1 GB 8GB GPU Commercial OK 2026-01-28
DeepSeek-OCR-2DeepSeek Vision-language 3.39B MoE 8K 8.5 GB5.0 GB3.3 GB 8GB GPU Commercial OK 2026-01-27
LongCat-Flash-LiteMeituan LongCat Text LLM 68.5B / 3B 320K 151 GB80.7 GB46.3 GB 2 × RTX 5090 (64GB) Commercial OK 2026-01-27
LongCat-Flash-Thinking-ZigZagMeituan LongCat · reasoning Text LLM 561.9B MoE 960K 1.24 TB657 GB376 GB 8 × H100 (640GB) Commercial OK 2026-01-23
Mistral-Small-4-119B-2603Mistral / Magistral / Devstral Text LLM 119B / 6.5B 1M 263 GB140 GB80.0 GB RTX PRO 6000 96GB Commercial OK 2026-01-23
Phi-4-reasoning-vision-15BPhi · reasoning Vision-language 15.1B 32K 35.4 GB19.9 GB12.3 GB 16GB GPU Commercial OK 2026-01-23
jina-embeddings-v5-text-nanoJina embeddings & rerankers Embedding 212M 8K 1.0 GB0.7 GB0.6 GB 8GB GPU Non-commercial 2026-01-22
jina-embeddings-v5-text-smallJina embeddings & rerankers Embedding 596M 32K 1.8 GB1.2 GB0.9 GB 8GB GPU Non-commercial 2026-01-22
Voxtral-Mini-4B-Realtime-2602Mistral / Magistral / Devstral Speech-to-text 4.43B 128K 10.2 GB5.7 GB3.5 GB 8GB GPU Commercial OK 2026-01-21
VibeVoice-ASRMicrosoft VibeVoice / Fara / Harrier Speech-to-text 8.67B — 19.6 GB10.6 GB6.3 GB 8GB GPU Commercial OK 2026-01-21
Qwen3-TTS-12Hz-0.6B-BaseQwen Text-to-speech / audio 915M — 2.5 GB1.6 GB1.1 GB 8GB GPU Commercial OK 2026-01-21
Qwen3-TTS-12Hz-1.7B-BaseQwen Text-to-speech / audio 1.93B — 4.7 GB2.8 GB1.8 GB 8GB GPU Commercial OK 2026-01-21
Qwen3-TTS-12Hz-0.6B-CustomVoiceQwen Text-to-speech / audio 906M — 2.5 GB1.6 GB1.1 GB 8GB GPU Commercial OK 2026-01-21
Qwen3-TTS-12Hz-1.7B-CustomVoiceQwen Text-to-speech / audio 1.92B — 4.7 GB2.7 GB1.8 GB 8GB GPU Commercial OK 2026-01-21
Qwen3-TTS-12Hz-1.7B-VoiceDesignQwen Text-to-speech / audio 1.92B — 4.7 GB2.7 GB1.8 GB 8GB GPU Commercial OK 2026-01-21
LFM2.5-1.2B-ThinkingLiquid LFM · reasoning Text LLM 1.17B 125K 3.2 GB2.0 GB1.4 GB 8GB GPU Custom licence – read it 2026-01-20
GLM-4.7-FlashGLM / CogVideoX (Z.ai) Text LLM 31.2B / 3B 198K 69.6 GB37.4 GB21.8 GB RTX 5090 32GB Commercial OK 2026-01-19
nemotron-colembed-vl-4b-v2NVIDIA Nemotron Embedding 4.83B 256K 11.1 GB6.1 GB3.7 GB 8GB GPU Non-commercial 2026-01-15
nemotron-colembed-vl-8b-v2NVIDIA Nemotron Embedding 8.77B 256K 19.8 GB10.7 GB6.3 GB 8GB GPU Non-commercial 2026-01-15
FLUX.2-klein-4BFLUX Image generation 3.88B — 9.0 GB5.0 GB3.1 GB 8GB GPU Commercial OK 2026-01-14
FLUX.2-klein-9BFLUX Image generation 9.08B — 20.5 GB11.1 GB6.6 GB 8GB GPU Non-commercial 2026-01-14
LongCat-Flash-Thinking-2601Meituan LongCat · reasoning Text LLM 560B / 27B 960K 1.23 TB655 GB374 GB 8 × H100 (640GB) Commercial OK 2026-01-14
LongCat-HeavyMode-SummaryMeituan LongCat Text LLM 561.9B MoE 960K 1.24 TB657 GB376 GB 8 × H100 (640GB) Commercial OK 2026-01-14
llama-nemotron-colembed-vl-3b-v2NVIDIA Nemotron Embedding 4.41B 128K 10.2 GB5.6 GB3.4 GB 8GB GPU Non-commercial 2026-01-14
Step-Audio-R1.1StepFun Step · reasoning Omni / audio LLM 33.5B 64K 76.3 GB41.8 GB25.0 GB RTX 5090 32GB Commercial OK 2026-01-14
Baichuan-M3-235BBaichuan Text LLM 235.1B MoE 40K 519 GB277 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2026-01-13
Step3-VL-10BStepFun Step Vision-language 10.2B 64K 24.1 GB13.6 GB8.5 GB 12GB GPU Commercial OK 2026-01-13
translategemma-4b-itGemma Translation 4.3B — 10.0 GB5.5 GB3.4 GB 8GB GPU Commercial with conditions 2026-01-12
translategemma-12b-itGemma Translation 12.2B — 27.3 GB14.7 GB8.6 GB 12GB GPU Commercial with conditions 2026-01-12
translategemma-27b-itGemma Translation 27.4B — 60.9 GB32.6 GB18.8 GB RTX 4090 24GB Commercial with conditions 2026-01-12
GLM-ImageGLM / CogVideoX (Z.ai) Image generation 6.93B — 15.7 GB8.6 GB5.1 GB 8GB GPU Commercial OK 2026-01-08
medgemma-1.5-4b-itGemma Vision-language 4.3B — 10.0 GB5.5 GB3.4 GB 8GB GPU Commercial with conditions 2026-01-07
Qwen3-VL-Embedding-2BQwen Embedding 2.13B 256K 5.2 GB3.0 GB1.9 GB 8GB GPU Commercial OK 2026-01-07
Qwen3-VL-Embedding-8BQwen Embedding 8.14B 256K 18.4 GB10.0 GB5.9 GB 8GB GPU Commercial OK 2026-01-07
Qwen3-VL-Reranker-2BQwen Reranker 2.13B 256K 5.2 GB3.0 GB1.9 GB 8GB GPU Commercial OK 2026-01-07
Qwen3-VL-Reranker-8BQwen Reranker 8.77B 256K 19.8 GB10.7 GB6.3 GB 8GB GPU Commercial OK 2026-01-07
LFM2.5-1.2B-InstructLiquid LFM Text LLM 1.17B 125K 3.2 GB2.0 GB1.4 GB 8GB GPU Custom licence – read it 2026-01-06
LFM2.5-VL-1.6BLiquid LFM Vision-language 1.6B 125K 4.1 GB2.5 GB1.7 GB 8GB GPU Custom licence – read it 2026-01-05
LFM2.5-1.2B-JPLiquid LFM Text LLM 1.17B 125K 3.2 GB2.0 GB1.4 GB 8GB GPU Custom licence – read it 2026-01-04
LTX-2LTX Video Video generation 18.9B — 42.0 GB22.6 GB13.1 GB 16GB GPU Custom licence – read it 2026-01-03
Kimi-K2.5Kimi Vision-language 1T / 32B 256K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial with conditions 2026-01-01
Qwen-Image-2512Qwen Image generation 20.4B — 45.4 GB24.4 GB14.1 GB 16GB GPU Commercial OK 2025-12-30
HY-MT1.5-1.8BTencent Hunyuan / Hy Translation 2.04B 256K 5.5 GB3.4 GB2.4 GB 8GB GPU Licence not stated 2025-12-25
GLM-4.7GLM / CogVideoX (Z.ai) Text LLM 358.3B MoE 198K 792 GB422 GB243 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-12-22
MiniMax-M2.1MiniMax Text LLM 228.7B MoE 192K 506 GB270 GB155 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2025-12-20
LFM2.5-Audio-1.5BLiquid LFM Text-to-speech / audio 1.47B — 3.7 GB2.2 GB1.5 GB 8GB GPU Custom licence – read it 2025-12-18
Qwen-Image-Edit-2511Qwen Image generation 20.4B — 45.4 GB24.4 GB14.1 GB 16GB GPU Commercial OK 2025-12-17
Qwen-Image-LayeredQwen Image generation 20.4B — 45.4 GB24.4 GB14.1 GB 16GB GPU Commercial OK 2025-12-17
MiMo-V2-FlashXiaomi MiMo Text LLM 309B / 15B 256K 680 GB362 GB207 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-12-16
Fun-ASR-MLT-Nano-2512FunAudioLLM speech Speech-to-text — — ——— — Commercial OK 2025-12-15
Molmo2-4BAi2 OLMo / Molmo Vision-language 4.85B 36K 12.4 GB7.4 GB4.9 GB 8GB GPU Commercial OK 2025-12-14
Molmo2-8BAi2 OLMo / Molmo Vision-language 8.66B 36K 20.8 GB11.8 GB7.5 GB 12GB GPU Commercial OK 2025-12-14
Molmo2-O-7BAi2 OLMo / Molmo Vision-language 7.76B 64K 21.9 GB13.9 GB10.0 GB 12GB GPU Commercial OK 2025-12-14
Fun-CosyVoice3-0.5B-2512FunAudioLLM speech Text-to-speech / audio 500M — 1.6 GB1.1 GB0.8 GB 8GB GPU Commercial OK 2025-12-11
granitelib-core-r1.0IBM Granite · reasoning Text LLM — — ——— — Commercial OK 2025-12-10
Olmo-3.1-32B-InstructAi2 OLMo / Molmo Text LLM 32.2B 64K 72.8 GB39.5 GB23.3 GB RTX 5090 32GB Commercial OK 2025-12-10
Olmo-3.1-32B-ThinkAi2 OLMo / Molmo · reasoning Text LLM 32.2B 64K 72.8 GB39.5 GB23.3 GB RTX 5090 32GB Commercial OK 2025-12-10
GLM-ASR-Nano-2512GLM / CogVideoX (Z.ai) Speech-to-text 2.26B 8K 5.5 GB3.1 GB2.0 GB 8GB GPU Commercial OK 2025-12-09
AutoGLM-Phone-9BGLM / CogVideoX (Z.ai) Vision-language 9B 64K 20.6 GB11.4 GB6.8 GB 8GB GPU Commercial OK 2025-12-08
Nemotron-Cascade-8B-ThinkingNVIDIA Nemotron · reasoning Text LLM 8.19B 32K 19.7 GB11.3 GB7.2 GB 8GB GPU Commercial with conditions 2025-12-08
Nemotron-Cascade-14B-ThinkingNVIDIA Nemotron · reasoning Text LLM 14.8B 32K 34.3 GB19.1 GB11.7 GB 16GB GPU Commercial with conditions 2025-12-08
GLM-4.6VGLM / CogVideoX (Z.ai) Vision-language 107.7B MoE 128K 239 GB128 GB73.9 GB RTX PRO 6000 96GB Commercial OK 2025-12-07
GLM-4.6V-FlashGLM / CogVideoX (Z.ai) Vision-language 10.3B 128K 23.5 GB12.9 GB7.7 GB 12GB GPU Commercial OK 2025-12-07
LongCat-Image-EditMeituan LongCat Image generation — — ——— — Commercial OK 2025-12-05
VoxCPM1.5MiniCPM / VoxCPM Text-to-speech / audio 802M — 2.3 GB1.4 GB1.0 GB 8GB GPU Commercial OK 2025-12-05
LongCat-ImageMeituan LongCat Image generation — — ——— — Commercial OK 2025-12-04
LongCat-Image-DevMeituan LongCat Image generation — — ——— — Commercial OK 2025-12-04
VibeVoice-Realtime-0.5BMicrosoft VibeVoice / Fara / Harrier Text-to-speech / audio 1.02B — 2.7 GB1.7 GB1.2 GB 8GB GPU Commercial OK 2025-12-04
NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA Nemotron Text LLM 31.6B / 3B 256K 70.0 GB37.5 GB21.6 GB RTX 4090 24GB Commercial with conditions 2025-12-04
llama-nemotron-rerank-vl-1b-v2NVIDIA Nemotron Reranker 1.68B 128K 4.2 GB2.5 GB1.6 GB 8GB GPU Commercial with conditions 2025-12-04
llama-nemotron-embed-vl-1b-v2NVIDIA Nemotron Embedding 1.68B 128K 4.2 GB2.5 GB1.6 GB 8GB GPU Commercial with conditions 2025-12-03
DeepSeek-V3.2DeepSeek Text LLM 685.4B MoE 160K 1.51 TB802 GB458 GB 8 × H100 (640GB) Commercial OK 2025-12-01
DeepSeek-V3.2-SpecialeDeepSeek Text LLM 685.4B MoE 160K 1.51 TB802 GB458 GB 8 × H100 (640GB) Commercial OK 2025-11-28
Devstral-2-123B-Instruct-2512Mistral / Magistral / Devstral · coding Text LLM 125B 256K 279 GB150 GB86.8 GB H200 141GB Custom licence – read it 2025-11-28
Mistral-Large-3-675B-Instruct-2512Mistral / Magistral / Devstral Text LLM 675B / 41B 288K 1.49 TB790 GB451 GB 8 × H100 (640GB) Commercial OK 2025-11-28
Nemotron-Orchestrator-8BNVIDIA Nemotron Text LLM 8.19B 40K 19.7 GB11.3 GB7.2 GB 8GB GPU Licence not stated 2025-11-25
FLUX.2-devFLUX Image generation 32.2B — 71.4 GB38.2 GB22.0 GB RTX 5090 32GB Non-commercial 2025-11-22
Olmo-3-7B-InstructAi2 OLMo / Molmo Text LLM 7.3B 64K 19.2 GB11.7 GB8.0 GB 12GB GPU Commercial OK 2025-11-19
Olmo-3-32B-ThinkAi2 OLMo / Molmo · reasoning Text LLM 32.2B 64K 72.8 GB39.5 GB23.3 GB RTX 5090 32GB Commercial OK 2025-11-19
HunyuanOCRTencent Hunyuan / Hy Vision-language 1.12B 128K 3.8 GB2.6 GB2.1 GB 8GB GPU Commercial with conditions 2025-11-18
HunyuanVideo-1.5Tencent Hunyuan / Hy Video generation — — ——— — Commercial with conditions 2025-11-18
Olmo-3-7B-ThinkAi2 OLMo / Molmo · reasoning Text LLM 7.3B 64K 19.2 GB11.7 GB8.0 GB 12GB GPU Commercial OK 2025-11-18
Voxtral-4B-TTS-2603Mistral / Magistral / Devstral Text-to-speech / audio 4B 125K 9.3 GB5.2 GB3.2 GB 8GB GPU Non-commercial 2025-11-17
NVIDIA-Nemotron-Parse-v1.1NVIDIA Nemotron Vision-language 957M — 2.6 GB1.6 GB1.1 GB 8GB GPU Commercial with conditions 2025-11-15
Wan2.2-Animate-14B-DiffusersWan Video generation 17.3B — 38.5 GB20.7 GB12.0 GB 16GB GPU Commercial OK 2025-11-12
ERNIE-4.5-VL-28B-A3B-ThinkingBaidu ERNIE · reasoning Vision-language 29.7B / 3B 128K 66.2 GB35.6 GB20.8 GB RTX 4090 24GB Commercial OK 2025-11-07
Nemotron-Flash-1BNVIDIA Nemotron Text LLM 965M MoE 35K 3.0 GB2.0 GB1.5 GB 8GB GPU Non-commercial 2025-11-06
Kimi-K2-ThinkingKimi · reasoning Text LLM 1T / 32B 256K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial with conditions 2025-11-04
Olmo-3-1125-32BAi2 OLMo / Molmo Text LLM 32.2B 64K 72.8 GB39.5 GB23.3 GB RTX 5090 32GB Commercial OK 2025-11-04
Ministral-3-3B-Instruct-2512-BF16Mistral / Magistral / Devstral Text LLM 4.25B 256K 10.7 GB6.3 GB4.2 GB 8GB GPU Commercial OK 2025-10-31
Ministral-3-8B-Instruct-2512-BF16Mistral / Magistral / Devstral Text LLM 8.92B 256K 21.3 GB12.1 GB7.6 GB 12GB GPU Commercial OK 2025-10-31
Ministral-3-14B-Instruct-2512-BF16Mistral / Magistral / Devstral Text LLM 13.9B 256K 32.5 GB18.1 GB11.1 GB 16GB GPU Commercial OK 2025-10-31
Ministral-3-3B-Reasoning-2512Mistral / Magistral / Devstral · reasoning Text LLM 4.25B 256K 10.7 GB6.3 GB4.2 GB 8GB GPU Commercial OK 2025-10-31
Ministral-3-8B-Reasoning-2512Mistral / Magistral / Devstral · reasoning Text LLM 8.92B 256K 21.3 GB12.1 GB7.6 GB 12GB GPU Commercial OK 2025-10-31
Ministral-3-14B-Reasoning-2512Mistral / Magistral / Devstral · reasoning Text LLM 13.9B 256K 32.5 GB18.1 GB11.1 GB 16GB GPU Commercial OK 2025-10-31
Kimi-Linear-48B-A3B-InstructKimi Text LLM 49.1B / 3B — 109 GB58.0 GB33.3 GB 2 × RTX 5090 (64GB) Commercial OK 2025-10-30
Fara-7BMicrosoft VibeVoice / Fara / Harrier Vision-language 8.29B 125K 19.2 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2025-10-30
Step-Audio-EditXStepFun Step Omni / audio LLM 3.53B — 11.5 GB7.8 GB6.1 GB 8GB GPU Licence not stated 2025-10-29
LFM2-ColBERT-350MLiquid LFM Embedding 353M 125K 1.3 GB0.9 GB0.7 GB 8GB GPU Custom licence – read it 2025-10-28
t5gemma-2-270m-270mGemma Vision-language 786M — 2.2 GB1.4 GB1.0 GB 8GB GPU Commercial with conditions 2025-10-25
t5gemma-2-1b-1bGemma Vision-language 2.12B — 5.2 GB3.0 GB1.9 GB 8GB GPU Commercial with conditions 2025-10-25
t5gemma-2-4b-4bGemma Vision-language 7.51B — 17.0 GB9.3 GB5.5 GB 8GB GPU Commercial with conditions 2025-10-25
LongCat-VideoMeituan LongCat Video generation — — ——— — Commercial OK 2025-10-24
LongCat-Flash-OmniMeituan LongCat Omni / audio LLM 560B / 27B 128K 1.23 TB655 GB374 GB 8 × H100 (640GB) Commercial OK 2025-10-23
LFM2-VL-3BLiquid LFM Vision-language 3B 125K 7.2 GB4.1 GB2.6 GB 8GB GPU Custom licence – read it 2025-10-22
MiniMax-M2MiniMax Text LLM 230B / 10B 192K 509 GB271 GB156 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2025-10-22
NVIDIA-Nemotron-Nano-12B-v2-VL-BF16NVIDIA Nemotron Vision-language 13.2B 128K 29.7 GB16.1 GB9.5 GB 12GB GPU Commercial with conditions 2025-10-21
Qwen3-VL-2B-InstructQwen Vision-language 2.13B 256K 6.1 GB3.9 GB2.9 GB 8GB GPU Commercial OK 2025-10-19
Qwen3-VL-32B-InstructQwen Vision-language 33.4B 256K 76.0 GB41.6 GB24.9 GB RTX 5090 32GB Commercial OK 2025-10-19
Qwen3-VL-2B-ThinkingQwen · reasoning Vision-language 2.13B 256K 6.1 GB3.9 GB2.9 GB 8GB GPU Commercial OK 2025-10-19
Qwen3-VL-32B-ThinkingQwen · reasoning Vision-language 33.4B 256K 76.0 GB41.6 GB24.9 GB RTX 5090 32GB Commercial OK 2025-10-19
DeepSeek-OCRDeepSeek Vision-language 3.34B MoE 8K 8.3 GB4.9 GB3.2 GB 8GB GPU Commercial OK 2025-10-17
llama-nemotron-embed-1b-v2NVIDIA Nemotron Embedding 1.24B 128K 3.2 GB1.9 GB1.3 GB 8GB GPU Commercial with conditions 2025-10-16
llama-nemotron-rerank-1b-v2NVIDIA Nemotron Reranker 1.24B 128K 3.2 GB1.9 GB1.3 GB 8GB GPU Commercial OK 2025-10-16
multitalker-parakeet-streaming-0.6b-v1NVIDIA Parakeet / Canary speech Speech-to-text 600M — 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial with conditions 2025-10-15
Ming-flash-omni-PreviewAnt Ling / Ring / Ming Omni / audio LLM 104.4B / 6B 32K 231 GB123 GB70.6 GB H100 80GB Commercial OK 2025-10-14
Qwen3-VL-4B-InstructQwen Vision-language 4.44B 256K 11.5 GB6.9 GB4.7 GB 8GB GPU Commercial OK 2025-10-11
Qwen3-VL-8B-InstructQwen Vision-language 8.77B 256K 21.0 GB12.0 GB7.6 GB 12GB GPU Commercial OK 2025-10-11
Qwen3-VL-4B-ThinkingQwen · reasoning Vision-language 4.44B 256K 11.5 GB6.9 GB4.7 GB 8GB GPU Commercial OK 2025-10-11
Qwen3-VL-8B-ThinkingQwen · reasoning Vision-language 8.77B 256K 21.0 GB12.0 GB7.6 GB 12GB GPU Commercial OK 2025-10-11
Ring-1TAnt Ling / Ring / Ming · reasoning Text LLM 1T / 50B 64K 2.20 TB1.17 TB670 GB 8 × H200 (1,128GB) Commercial OK 2025-10-10
functiongemma-270m-itGemma Text LLM 268M — 1.1 GB0.8 GB0.7 GB 8GB GPU Commercial with conditions 2025-10-08
granite-4.0-350mIBM Granite Text LLM 352M 32K 1.5 GB1.1 GB1.0 GB 8GB GPU Commercial OK 2025-10-07
granite-4.0-1bIBM Granite Text LLM 1.63B 128K 4.8 GB3.1 GB2.3 GB 8GB GPU Commercial OK 2025-10-07
granite-4.0-h-350mIBM Granite Text LLM 340M 32K 1.3 GB0.9 GB0.8 GB 8GB GPU Commercial OK 2025-10-07
granite-4.0-h-1bIBM Granite Text LLM 1.46B 128K 3.8 GB2.3 GB1.5 GB 8GB GPU Commercial OK 2025-10-07
LFM2-8B-A1BLiquid LFM Text LLM 8.34B / 1B 125K 18.9 GB10.3 GB6.2 GB 8GB GPU Custom licence – read it 2025-10-07
llama-embed-nemotron-8bNVIDIA Nemotron Embedding 7.5B 128K 17.0 GB9.3 GB5.5 GB 8GB GPU Non-commercial 2025-10-07
olmOCR-2-7B-1025Ai2 OLMo / Molmo Vision-language 8.29B 125K 19.2 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2025-10-06
Ling-1TAnt Ling / Ring / Ming Text LLM 1T / 50B 32K 2.20 TB1.17 TB670 GB 8 × H200 (1,128GB) Commercial OK 2025-10-02
omni-embed-nemotron-3bNVIDIA Nemotron Embedding 4.7B 32K 10.8 GB6.0 GB3.6 GB 8GB GPU Non-commercial 2025-09-30
Qwen3-VL-30B-A3B-InstructQwen Vision-language 31.1B / 3B 256K 69.7 GB37.6 GB22.0 GB RTX 5090 32GB Commercial OK 2025-09-30
Qwen3-VL-30B-A3B-ThinkingQwen · reasoning Vision-language 31.1B / 3B 256K 69.7 GB37.6 GB22.0 GB RTX 5090 32GB Commercial OK 2025-09-30
DeepSeek-V3.2-Exp-BaseDeepSeek Text LLM 685.4B MoE 160K 1.51 TB802 GB458 GB 8 × H100 (640GB) Commercial OK 2025-09-29
GLM-4.6GLM / CogVideoX (Z.ai) Text LLM 356.8B MoE 198K 789 GB421 GB242 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-09-29
HunyuanImage-3.0-InstructTencent Hunyuan / Hy Image generation 80B / 13B 22K 177 GB94.0 GB53.9 GB 2 × RTX 5090 (64GB) Custom licence – read it 2025-09-25
Ring-flash-linear-2.0Ant Ling / Ring / Ming · reasoning Text LLM 104.2B / 6.1B 128K 230 GB123 GB70.5 GB H100 80GB Commercial OK 2025-09-25
Ring-mini-linear-2.0Ant Ling / Ring / Ming · reasoning Text LLM 16.4B / 1.6B 128K 36.9 GB20.0 GB11.8 GB 16GB GPU Commercial OK 2025-09-24
Hunyuan3D-PartTencent Hunyuan / Hy Text LLM — — ——— — Licence not stated 2025-09-23
Qwen3Guard-Gen-0.6BQwen Safety / guard 752M 32K 3.1 GB2.3 GB1.9 GB 8GB GPU Commercial OK 2025-09-23
Qwen3Guard-Gen-4BQwen Safety / guard 4.41B 32K 11.4 GB6.9 GB4.7 GB 8GB GPU Commercial OK 2025-09-23
Qwen3Guard-Gen-8BQwen Safety / guard 8.19B 32K 19.7 GB11.3 GB7.2 GB 8GB GPU Commercial OK 2025-09-23
Qwen3Guard-Stream-0.6BQwen Safety / guard 597M 8K 2.8 GB2.1 GB1.8 GB 8GB GPU Commercial OK 2025-09-23
LFM2-2.6BLiquid LFM Text LLM 2.57B 125K 6.3 GB3.6 GB2.3 GB 8GB GPU Custom licence – read it 2025-09-22
Qwen-Image-Edit-2509Qwen Image generation 20.4B — 45.4 GB24.4 GB14.1 GB 16GB GPU Commercial OK 2025-09-22
Qwen3-VL-235B-A22B-InstructQwen Vision-language 235.7B / 22B 256K 521 GB278 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-09-22
Qwen3-VL-235B-A22B-ThinkingQwen · reasoning Vision-language 235.7B / 22B 256K 521 GB278 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-09-22
LongCat-Flash-ThinkingMeituan LongCat · reasoning Text LLM 560B / 27B 128K 1.23 TB655 GB374 GB 8 × H100 (640GB) Commercial OK 2025-09-21
Qwen3-Omni-30B-A3B-InstructQwen Omni / audio LLM 35.3B / 3B 64K 78.9 GB42.5 GB24.8 GB RTX 5090 32GB Commercial OK 2025-09-20
Ring-flash-2.0Ant Ling / Ring / Ming · reasoning Text LLM 100B / 6.1B 32K 221 GB118 GB67.7 GB H100 80GB Commercial OK 2025-09-19
gpt-oss-safeguard-20bgpt-oss Safety / guard 21B / 3.6B 128K 46.9 GB25.2 GB14.7 GB RTX 4090 24GB Commercial OK 2025-09-18
gpt-oss-safeguard-120bgpt-oss Safety / guard 117B / 5.1B 128K 258 GB138 GB78.8 GB RTX PRO 6000 96GB Commercial OK 2025-09-18
jina-reranker-v3Jina embeddings & rerankers Reranker 597M 128K 1.8 GB1.2 GB0.9 GB 8GB GPU Non-commercial 2025-09-18
MiMo-Audio-7B-InstructXiaomi MiMo Omni / audio LLM 8.02B 8K 19.4 GB11.1 GB7.1 GB 8GB GPU Commercial OK 2025-09-18
Ling-flash-2.0Ant Ling / Ring / Ming Text LLM 100B / 6.1B 32K 221 GB118 GB67.7 GB H100 80GB Commercial OK 2025-09-17
granite-4.0-h-microIBM Granite Text LLM 3.19B 128K 7.6 GB4.3 GB2.7 GB 8GB GPU Commercial OK 2025-09-16
granite-4.0-h-smallIBM Granite Text LLM 32.2B MoE 128K 71.5 GB38.3 GB22.1 GB RTX 5090 32GB Commercial OK 2025-09-16
granite-4.0-h-tinyIBM Granite Text LLM 6.94B MoE 128K 15.8 GB8.7 GB5.2 GB 8GB GPU Commercial OK 2025-09-16
granite-4.0-microIBM Granite Text LLM 3.4B 128K 8.7 GB5.1 GB3.4 GB 8GB GPU Commercial OK 2025-09-16
VoxCPM-0.5BMiniCPM / VoxCPM Text-to-speech / audio 500M — 1.6 GB1.1 GB0.8 GB 8GB GPU Commercial OK 2025-09-16
Qwen3-Omni-30B-A3B-ThinkingQwen · reasoning Omni / audio LLM 31.7B / 3B 64K 71.1 GB38.4 GB22.5 GB RTX 5090 32GB Commercial OK 2025-09-15
Magistral-Small-2509Mistral / Magistral / Devstral · reasoning Text LLM 24B 128K 54.7 GB29.9 GB17.9 GB RTX 4090 24GB Commercial OK 2025-09-12
Olmo-3-1025-7BAi2 OLMo / Molmo Text LLM 7.3B 64K 19.2 GB11.7 GB8.0 GB 12GB GPU Commercial OK 2025-09-12
Qwen3-Next-80B-A3B-InstructQwen Text LLM 81.3B / 3B 256K 180 GB95.7 GB54.9 GB 2 × RTX 5090 (64GB) Commercial OK 2025-09-09
Qwen3-Next-80B-A3B-ThinkingQwen · reasoning Text LLM 81.3B / 3B 256K 180 GB95.7 GB54.9 GB 2 × RTX 5090 (64GB) Commercial OK 2025-09-09
ERNIE-4.5-21B-A3B-ThinkingBaidu ERNIE · reasoning Text LLM 21.8B / 3B 128K 49.0 GB26.5 GB15.5 GB RTX 4090 24GB Commercial OK 2025-09-08
Ling-mini-2.0Ant Ling / Ring / Ming Text LLM 16.3B / 1.43B 32K 36.6 GB19.8 GB11.7 GB 16GB GPU Commercial OK 2025-09-08
vaultgemma-1bGemma Text LLM 1.04B — 2.8 GB1.7 GB1.2 GB 8GB GPU Commercial with conditions 2025-09-05
Kimi-K2-Instruct-0905Kimi Text LLM 1T / 32B 256K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial with conditions 2025-09-03
MiniCPM4.1-8BMiniCPM / VoxCPM Text LLM 8.19B 64K 18.8 GB10.3 GB6.2 GB 8GB GPU Commercial OK 2025-09-02
LongCat-Flash-ChatMeituan LongCat Text LLM 560B / 27B 128K 1.23 TB655 GB374 GB 8 × H100 (640GB) Commercial OK 2025-08-29
Hunyuan-MT-7BTencent Hunyuan / Hy Translation 8.03B 32K 19.2 GB11.0 GB6.9 GB 8GB GPU Licence not stated 2025-08-28
Step-Audio-2-miniStepFun Step Omni / audio LLM 8.31B 16K 19.3 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2025-08-28
jina-code-embeddings-1.5bJina embeddings & rerankers · coding Embedding 1.54B 32K 3.9 GB2.3 GB1.5 GB 8GB GPU Non-commercial 2025-08-26
VibeVoice-1.5BMicrosoft VibeVoice / Fara / Harrier Text-to-speech / audio 2.7B — 6.4 GB3.7 GB2.3 GB 8GB GPU Commercial OK 2025-08-25
Wan2.2-S2V-14BWan Video generation 16.3B — 36.4 GB19.5 GB11.4 GB 16GB GPU Commercial OK 2025-08-25
MiniCPM-V-4_5MiniCPM / VoxCPM Vision-language 8.7B 40K 20.8 GB11.9 GB7.5 GB 12GB GPU Commercial OK 2025-08-24
grok-2xAI Grok Text LLM — MoE 128K ——— — Licence not stated 2025-08-22
NVIDIA-Nemotron-Nano-12B-v2NVIDIA Nemotron Text LLM 12.3B 128K 27.8 GB15.1 GB8.9 GB 12GB GPU Commercial with conditions 2025-08-21
Seed-OSS-36B-InstructByteDance Seed Text LLM 36.2B 512K 82.2 GB44.9 GB26.8 GB RTX 5090 32GB Commercial OK 2025-08-20
DeepSeek-V3.1-BaseDeepSeek Text LLM 671B / 37B 160K 1.48 TB785 GB449 GB 8 × H100 (640GB) Commercial OK 2025-08-19
jina-code-embeddings-0.5bJina embeddings & rerankers · coding Embedding 494M 32K 1.6 GB1.1 GB0.8 GB 8GB GPU Non-commercial 2025-08-19
Intern-S1-miniInternLM / Intern-S Vision-language 8.54B 64K 20.5 GB11.7 GB7.4 GB 12GB GPU Commercial OK 2025-08-18
Qwen-Image-EditQwen Image generation 20.4B — 45.4 GB24.4 GB14.1 GB 16GB GPU Commercial OK 2025-08-17
command-a-reasoning-08-2025Cohere Command / Aya · reasoning Text LLM 111.1B — 245 GB130 GB74.6 GB RTX PRO 6000 96GB Non-commercial 2025-08-12
LFM2-VL-450MLiquid LFM Vision-language 451M 125K 1.6 GB1.1 GB0.9 GB 8GB GPU Custom licence – read it 2025-08-12
LFM2-VL-1.6BLiquid LFM Vision-language 1.58B 125K 4.1 GB2.5 GB1.7 GB 8GB GPU Custom licence – read it 2025-08-12
NVIDIA-Nemotron-Nano-9B-v2NVIDIA Nemotron Text LLM 8.89B 128K 20.2 GB11.0 GB6.6 GB 8GB GPU Commercial with conditions 2025-08-12
Baichuan-M2-32BBaichuan Text LLM 32.8B 128K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2025-08-10
GLM-4.5VGLM / CogVideoX (Z.ai) Vision-language 107.7B MoE 64K 239 GB128 GB73.9 GB RTX PRO 6000 96GB Commercial OK 2025-08-10
MiMo-VL-7B-RL-2508Xiaomi MiMo · reasoning Vision-language 8.31B 125K 20.0 GB11.4 GB7.2 GB 8GB GPU Commercial OK 2025-08-07
Qwen3-4B-Instruct-2507Qwen Text LLM 4.02B 256K 10.6 GB6.4 GB4.4 GB 8GB GPU Commercial OK 2025-08-05
Qwen3-4B-Thinking-2507Qwen · reasoning Text LLM 4.02B 256K 10.6 GB6.4 GB4.4 GB 8GB GPU Commercial OK 2025-08-05
gpt-oss-20bgpt-oss Text LLM 21B / 3.6B 128K 46.9 GB25.2 GB14.7 GB RTX 4090 24GB Commercial OK 2025-08-04
gpt-oss-120bgpt-oss Text LLM 117B / 5.1B 128K 258 GB138 GB78.8 GB RTX PRO 6000 96GB Commercial OK 2025-08-04
granite-embedding-reranker-english-r2IBM Granite Reranker 150M 8K 0.8 GB0.7 GB0.6 GB 8GB GPU Commercial OK 2025-08-04
canary-1b-v2NVIDIA Parakeet / Canary speech Speech-to-text 979M — 2.7 GB1.6 GB1.2 GB 8GB GPU Commercial OK 2025-08-04
Qwen-ImageQwen Image generation 20.4B — 45.4 GB24.4 GB14.1 GB 16GB GPU Commercial OK 2025-08-02
Qwen3-Coder-30B-A3B-InstructQwen · coding Text LLM 30.5B / 3B 256K 68.5 GB37.0 GB21.7 GB RTX 5090 32GB Commercial OK 2025-07-31
gemma-3-270m-itGemma Text LLM 268M — 1.1 GB0.8 GB0.7 GB 8GB GPU Commercial with conditions 2025-07-30
Hunyuan-0.5B-InstructTencent Hunyuan / Hy Text LLM 539M 256K 2.5 GB1.9 GB1.7 GB 8GB GPU Licence not stated 2025-07-30
Hunyuan-1.8B-InstructTencent Hunyuan / Hy Text LLM 1.79B 256K 5.0 GB3.1 GB2.2 GB 8GB GPU Licence not stated 2025-07-30
Hunyuan-4B-InstructTencent Hunyuan / Hy Text LLM 4.22B 256K 11.0 GB6.6 GB4.5 GB 8GB GPU Licence not stated 2025-07-30
Hunyuan-7B-InstructTencent Hunyuan / Hy Text LLM 7.5B 32K 18.1 GB10.3 GB6.6 GB 8GB GPU Licence not stated 2025-07-30
CosyVoice2-0.5BFunAudioLLM speech Text-to-speech / audio 500M — 1.6 GB1.1 GB0.8 GB 8GB GPU Commercial OK 2025-07-29
Qwen3-30B-A3B-Thinking-2507Qwen · reasoning Text LLM 30.5B / 3B 256K 68.5 GB37.0 GB21.7 GB RTX 5090 32GB Commercial OK 2025-07-29
command-a-vision-07-2025Cohere Command / Aya Vision-language 111.9B — 247 GB131 GB75.1 GB RTX PRO 6000 96GB Non-commercial 2025-07-28
Qwen3-30B-A3B-Instruct-2507Qwen Text LLM 30.5B / 3B 256K 68.5 GB37.0 GB21.7 GB RTX 5090 32GB Commercial OK 2025-07-28
step3StepFun Step Vision-language 321B / 38B — 740 GB408 GB247 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-07-28
Wan2.2-I2V-A14B-DiffusersWan Video generation 14.3B / 14B — 31.9 GB17.2 GB10.0 GB 12GB GPU Commercial OK 2025-07-28
Wan2.2-T2V-A14B-DiffusersWan Video generation 14.3B / 14B — 31.9 GB17.2 GB10.0 GB 12GB GPU Commercial OK 2025-07-28
Wan2.2-TI2V-5B-DiffusersWan Video generation 5B — 11.5 GB6.3 GB3.8 GB 8GB GPU Commercial OK 2025-07-28
Llama-3_3-Nemotron-Super-49B-v1_5NVIDIA Nemotron Text LLM 49.9B 128K 112 GB60.4 GB35.4 GB 2 × RTX 5090 (64GB) Commercial with conditions 2025-07-25
Qwen3-235B-A22B-Thinking-2507Qwen · reasoning Text LLM 235.1B / 22B 256K 519 GB277 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-07-25
Intern-S1InternLM / Intern-S Vision-language 240.7B MoE 64K 532 GB283 GB163 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-07-24
Qwen3-Coder-480B-A35B-InstructQwen · coding Text LLM 480.2B / 35B 256K 1.06 TB564 GB323 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-07-22
Qwen3-235B-A22B-Instruct-2507Qwen Text LLM 235.1B / 22B 256K 519 GB277 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-07-21
GLM-4.5GLM / CogVideoX (Z.ai) Text LLM 355B / 32B 128K 785 GB419 GB240 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-07-20
GLM-4.5-AirGLM / CogVideoX (Z.ai) Text LLM 106B / 12B 128K 235 GB126 GB72.7 GB RTX PRO 6000 96GB Commercial OK 2025-07-20
Magistral-Small-2507Mistral / Magistral / Devstral · reasoning Text LLM 23.6B 40K 53.7 GB29.4 GB17.6 GB RTX 4090 24GB Commercial OK 2025-07-18
embeddinggemma-300mGemma Embedding 303M — 1.2 GB0.9 GB0.7 GB 8GB GPU Commercial with conditions 2025-07-17
granite-embedding-english-r2IBM Granite Embedding 149M 8K 0.8 GB0.7 GB0.6 GB 8GB GPU Commercial OK 2025-07-17
granite-embedding-small-english-r2IBM Granite Embedding 48M 8K 0.6 GB0.6 GB0.5 GB 8GB GPU Commercial OK 2025-07-17
LTX-Video-0.9.8-13B-distilledLTX Video Video generation 13B — 29.2 GB15.7 GB9.2 GB 12GB GPU Custom licence – read it 2025-07-17
Seed-X-Instruct-7BByteDance Seed Translation 7.51B 32K 18.1 GB10.4 GB6.6 GB 8GB GPU Commercial OK 2025-07-16
Seed-X-PPO-7BByteDance Seed Translation 7.51B 32K 18.1 GB10.4 GB6.6 GB 8GB GPU Commercial OK 2025-07-16
OpenReasoning-Nemotron-7BNVIDIA Nemotron · reasoning Text LLM 7.62B 128K 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2025-07-15
MiniCPM-V-4MiniCPM / VoxCPM Vision-language 4.06B 32K 9.7 GB5.5 GB3.5 GB 8GB GPU Commercial OK 2025-07-12
Kimi-K2-InstructKimi Text LLM 1T / 32B 128K 2.20 TB1.17 TB668 GB 8 × H200 (1,128GB) Commercial with conditions 2025-07-11
LFM2-350MLiquid LFM Text LLM 354M 125K 1.5 GB1.2 GB1.0 GB 8GB GPU Custom licence – read it 2025-07-10
LFM2-700MLiquid LFM Text LLM 742M 125K 2.4 GB1.6 GB1.3 GB 8GB GPU Custom licence – read it 2025-07-10
LFM2-1.2BLiquid LFM Text LLM 1.17B 125K 3.3 GB2.1 GB1.6 GB 8GB GPU Custom licence – read it 2025-07-10
medgemma-27b-itGemma Vision-language 27.4B — 60.9 GB32.6 GB18.8 GB RTX 4090 24GB Commercial with conditions 2025-07-09
SmolLM3-3BSmolLM / SmolVLM Text LLM 3.08B 64K 7.9 GB4.7 GB3.2 GB 8GB GPU Commercial OK 2025-07-08
FLUX.1-Krea-devFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Non-commercial 2025-07-07
Devstral-Small-2507Mistral / Magistral / Devstral · coding Text LLM 23.6B 128K 53.7 GB29.4 GB17.6 GB RTX 4090 24GB Commercial OK 2025-07-04
jina-embeddings-v4-vllm-retrievalJina embeddings & rerankers Embedding 3.75B 125K 8.8 GB4.9 GB3.0 GB 8GB GPU Licence not stated 2025-07-01
Voxtral-Mini-3B-2507Mistral / Magistral / Devstral Text LLM 4.68B 128K 11.8 GB7.0 GB4.6 GB 8GB GPU Commercial OK 2025-07-01
Voxtral-Small-24B-2507Mistral / Magistral / Devstral Omni / audio LLM 24.3B 128K 55.2 GB30.2 GB18.0 GB RTX 4090 24GB Commercial OK 2025-07-01
ERNIE-4.5-0.3B-PTBaidu ERNIE Text LLM 361M 128K 1.4 GB1.1 GB0.9 GB 8GB GPU Commercial OK 2025-06-28
ERNIE-4.5-21B-A3B-PTBaidu ERNIE Text LLM 21.9B / 3B 128K 49.3 GB26.6 GB15.6 GB RTX 4090 24GB Commercial OK 2025-06-28
ERNIE-4.5-300B-A47B-PTBaidu ERNIE Text LLM 300.5B / 47B 128K 663 GB354 GB203 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-28
ERNIE-4.5-300B-A47B-PaddleBaidu ERNIE Text LLM 300.5B / 47B 128K 663 GB354 GB203 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-28
ERNIE-4.5-VL-28B-A3B-PTBaidu ERNIE Vision-language 29.4B / 3B 128K 65.6 GB35.3 GB20.6 GB RTX 4090 24GB Commercial OK 2025-06-28
ERNIE-4.5-VL-424B-A47B-PTBaidu ERNIE Vision-language 423.5B / 47B 128K 934 GB497 GB285 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-28
ERNIE-4.5-VL-424B-A47B-PaddleBaidu ERNIE Vision-language 423.5B / 47B 128K 934 GB497 GB285 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-28
GLM-4.1V-9B-ThinkingGLM / CogVideoX (Z.ai) · reasoning Vision-language 10.3B 64K 23.5 GB12.9 GB7.7 GB 12GB GPU Commercial OK 2025-06-28
canary-qwen-2.5bNVIDIA Parakeet / Canary speech Speech-to-text 2.56B — 6.1 GB3.5 GB2.2 GB 8GB GPU Commercial OK 2025-06-26
Hunyuan-A13B-InstructTencent Hunyuan / Hy Text LLM 80.4B / 13B 32K 178 GB95.5 GB55.2 GB 2 × RTX 5090 (64GB) Commercial with conditions 2025-06-25
Phi-mini-MoE-instructPhi Text LLM 7.6B / 2.4B 4K 18.3 GB10.5 GB6.6 GB 8GB GPU Commercial OK 2025-06-23
Phi-tiny-MoE-instructPhi Text LLM 3.8B / 1.1B 4K 9.4 GB5.5 GB3.6 GB 8GB GPU Commercial OK 2025-06-23
Kimi-VL-A3B-Thinking-2506Kimi · reasoning Vision-language 16.4B / 3B 128K 36.8 GB19.9 GB11.7 GB 16GB GPU Commercial OK 2025-06-21
t5gemma-b-b-ul2Gemma Text LLM 591M — 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial with conditions 2025-06-19
t5gemma-s-s-prefixlmGemma Text LLM 313M — 1.2 GB0.9 GB0.7 GB 8GB GPU Commercial with conditions 2025-06-19
t5gemma-s-s-ul2Gemma Text LLM 313M — 1.2 GB0.9 GB0.7 GB 8GB GPU Commercial with conditions 2025-06-19
Mistral-Small-3.2-24B-Instruct-2506Mistral / Magistral / Devstral Text LLM 24B 128K 54.7 GB29.9 GB17.9 GB RTX 4090 24GB Commercial OK 2025-06-19
Phi-4-mini-flash-reasoningPhi · reasoning Text LLM 3.85B 256K 10.3 GB6.3 GB4.4 GB 8GB GPU Commercial OK 2025-06-19
Kimi-Dev-72BKimi Text LLM 72.7B 128K 163 GB88.2 GB51.7 GB 2 × RTX 5090 (64GB) Commercial OK 2025-06-16
MiniMax-M1-80kMiniMax Text LLM 456B / 45.9B — 1.00 TB534 GB305 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-13
gemma-3n-E2B-itGemma Vision-language 5.44B — 12.5 GB6.9 GB4.1 GB 8GB GPU Commercial with conditions 2025-06-12
FlexOlmo-7x7B-1TAi2 OLMo / Molmo Text LLM 1T MoE 4K 2.20 TB1.17 TB672 GB 8 × H200 (1,128GB) Commercial OK 2025-06-11
SeedVR-3BByteDance Seed Video generation 3B — 7.1 GB4.0 GB2.5 GB 8GB GPU Commercial OK 2025-06-11
SeedVR2-7BByteDance Seed Video generation 7B — 15.9 GB8.7 GB5.2 GB 8GB GPU Commercial OK 2025-06-11
Step-Audio-AQAAStepFun Step Omni / audio LLM 136.7B — 337 GB196 GB127 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-11
SeedVR2-3BByteDance Seed Video generation 3B — 7.1 GB4.0 GB2.5 GB 8GB GPU Commercial OK 2025-06-10
MiniCPM4-0.5BMiniCPM / VoxCPM Text LLM 434M 32K 1.6 GB1.1 GB0.9 GB 8GB GPU Commercial OK 2025-06-05
MiniCPM4-8BMiniCPM / VoxCPM Text LLM 8.19B 32K 18.8 GB10.3 GB6.2 GB 8GB GPU Commercial OK 2025-06-05
MiniCPM4-MCPMiniCPM / VoxCPM Text LLM 8.19B 32K 18.8 GB10.3 GB6.2 GB 8GB GPU Commercial OK 2025-06-05
MiniMax-M1-40kMiniMax Text LLM 456B / 45.9B — 1.00 TB534 GB305 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-06-05
Nemotron-H-8B-Reasoning-128KNVIDIA Nemotron · reasoning Text LLM 8.1B 128K 18.5 GB10.1 GB6.0 GB 8GB GPU Non-commercial 2025-06-05
Magistral-Small-2506Mistral / Magistral / Devstral · reasoning Text LLM 23.6B 40K 53.7 GB29.4 GB17.6 GB RTX 4090 24GB Commercial OK 2025-06-04
gemma-3n-E4B-itGemma Vision-language 7.85B — 17.8 GB9.7 GB5.7 GB 8GB GPU Commercial with conditions 2025-06-03
granite-guardian-3.3-8bIBM Granite Safety / guard 8.17B 128K 19.8 GB11.4 GB7.3 GB 12GB GPU Commercial OK 2025-06-03
granite-vision-3.3-2bIBM Granite Vision-language 2.98B 128K 7.7 GB4.6 GB3.2 GB 8GB GPU Commercial OK 2025-06-03
MiniMax-Text-01-hfMiniMax Text LLM 456B / 45.9B 4M 1.00 TB534 GB305 GB 4 × RTX PRO 6000 (384GB) Custom licence – read it 2025-06-03
Llama-3.1-Nemotron-Nano-VL-8B-V1NVIDIA Nemotron Vision-language 8.72B 128K 20.8 GB11.8 GB7.4 GB 12GB GPU Commercial with conditions 2025-06-03
Qwen3-Embedding-0.6BQwen Embedding 596M 32K 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2025-06-03
Qwen3-Embedding-4BQwen Embedding 4.02B 40K 9.3 GB5.2 GB3.2 GB 8GB GPU Commercial OK 2025-06-03
Qwen3-Embedding-8BQwen Embedding 7.57B 40K 17.1 GB9.3 GB5.5 GB 8GB GPU Commercial OK 2025-06-03
Qwen3-Reranker-4BQwen Reranker 4.02B 40K 9.3 GB5.2 GB3.2 GB 8GB GPU Commercial OK 2025-06-03
MiMo-VL-7B-RLXiaomi MiMo · reasoning Vision-language 8.31B 125K 20.0 GB11.4 GB7.2 GB 8GB GPU Commercial OK 2025-05-30
DeepSeek-R1-0528-Qwen3-8BDeepSeek · reasoning Text LLM 8.19B 128K 19.7 GB11.3 GB7.2 GB 8GB GPU Commercial OK 2025-05-29
Qwen3-Reranker-0.6BQwen Reranker 596M 40K 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2025-05-29
Qwen3-Reranker-8BQwen Reranker 8.19B 40K 18.5 GB10.1 GB6.0 GB 8GB GPU Commercial OK 2025-05-29
DeepSeek-R1-0528DeepSeek · reasoning Text LLM 684.5B MoE 160K 1.51 TB801 GB458 GB 8 × H100 (640GB) Commercial OK 2025-05-28
FLUX.1-Kontext-devFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Non-commercial 2025-05-28
medgemma-4b-itGemma Vision-language 4.3B — 10.0 GB5.5 GB3.4 GB 8GB GPU Commercial with conditions 2025-05-19
medgemma-27b-text-itGemma Text LLM 27B — 59.9 GB32.1 GB18.5 GB RTX 4090 24GB Commercial with conditions 2025-05-19
granite-docling-258MIBM Granite Vision-language 258M 8K 1.3 GB1.0 GB0.9 GB 8GB GPU Commercial OK 2025-05-19
BAGEL-7B-MoTByteDance Seed Omni / audio LLM 14.7B 32K 33.3 GB18.1 GB10.8 GB 12GB GPU Commercial OK 2025-05-19
bge-code-v1BGE embeddings & rerankers · coding Embedding 1.54B 32K 3.9 GB2.3 GB1.5 GB 8GB GPU Commercial OK 2025-05-15
LTX-Video-0.9.7-distilledLTX Video Video generation 13B — 29.2 GB15.7 GB9.2 GB 12GB GPU Custom licence – read it 2025-05-14
LTX-Video-0.9.7-devLTX Video Video generation 13B — 29.2 GB15.7 GB9.2 GB 12GB GPU Custom licence – read it 2025-05-13
Wan2.1-VACE-1.3BWan Video generation 2.15B — 5.2 GB3.0 GB1.9 GB 8GB GPU Commercial OK 2025-05-13
Wan2.1-VACE-14BWan Video generation 17.3B — 38.6 GB20.8 GB12.1 GB 16GB GPU Commercial OK 2025-05-13
Devstral-Small-2505Mistral / Magistral / Devstral · coding Text LLM 23.6B 128K 53.7 GB29.4 GB17.6 GB RTX 4090 24GB Commercial OK 2025-05-12
stable-audio-open-smallStable Diffusion / Stable Audio Text-to-speech / audio 497M — 1.6 GB1.1 GB0.8 GB 8GB GPU Commercial with conditions 2025-05-12
Ling-lite-1.5Ant Ling / Ring / Ming Text LLM 16.8B / 2.75B 32K 37.9 GB20.6 GB12.2 GB 16GB GPU Commercial OK 2025-05-11
jina-embeddings-v4Jina embeddings & rerankers Embedding 3.75B 125K 8.8 GB4.9 GB3.0 GB 8GB GPU Licence not stated 2025-05-07
Llama-3.1-Nemotron-Nano-4B-v1.1NVIDIA Nemotron Text LLM 4.51B 128K 11.5 GB6.8 GB4.6 GB 8GB GPU Commercial with conditions 2025-05-03
granite-4.0-tiny-previewIBM Granite Text LLM 6.67B MoE 128K 15.2 GB8.4 GB5.0 GB 8GB GPU Commercial OK 2025-04-30
Qwen2.5-Omni-3BQwen Omni / audio LLM 5.54B 32K 13.0 GB7.3 GB4.5 GB 8GB GPU Non-commercial 2025-04-30
MiMo-7B-BaseXiaomi MiMo Text LLM 7.83B 32K 18.9 GB10.9 GB6.9 GB 8GB GPU Commercial OK 2025-04-29
MiMo-7B-RLXiaomi MiMo · reasoning Text LLM 7.83B 32K 18.9 GB10.9 GB6.9 GB 8GB GPU Commercial OK 2025-04-29
Phi-4-mini-reasoningPhi · reasoning Text LLM 3.84B 128K 10.0 GB6.1 GB4.1 GB 8GB GPU Commercial OK 2025-04-29
granite-speech-3.3-2bIBM Granite Speech-to-text 3.01B 128K 7.1 GB4.0 GB2.5 GB 8GB GPU Commercial OK 2025-04-28
Llama-Prompt-Guard-2-22MLlama Safety / guard 71M — 0.7 GB0.6 GB0.5 GB 8GB GPU Commercial with conditions 2025-04-28
Llama-Prompt-Guard-2-86MLlama Safety / guard 279M — 1.1 GB0.8 GB0.7 GB 8GB GPU Commercial with conditions 2025-04-28
Qwen3-0.6BQwen Text LLM 752M 40K 3.1 GB2.3 GB1.9 GB 8GB GPU Commercial OK 2025-04-27
Qwen3-1.7BQwen Text LLM 2.03B 40K 5.9 GB3.8 GB2.8 GB 8GB GPU Commercial OK 2025-04-27
Qwen3-4BQwen Text LLM 4.02B 40K 10.6 GB6.4 GB4.4 GB 8GB GPU Commercial OK 2025-04-27
Qwen3-8BQwen Text LLM 8.19B 40K 19.7 GB11.3 GB7.2 GB 8GB GPU Commercial OK 2025-04-27
Qwen3-14BQwen Text LLM 14.8B 40K 34.3 GB19.1 GB11.7 GB 16GB GPU Commercial OK 2025-04-27
Qwen3-30B-A3BQwen Text LLM 30.5B / 3B 40K 68.5 GB37.0 GB21.7 GB RTX 5090 32GB Commercial OK 2025-04-27
Qwen3-32BQwen Text LLM 32.8B 40K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2025-04-27
Qwen3-235B-A22BQwen Text LLM 235.1B / 22B 40K 519 GB277 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-04-27
Seed-Coder-8B-InstructByteDance Seed · coding Text LLM 8.25B 32K 19.7 GB11.2 GB7.1 GB 8GB GPU Commercial OK 2025-04-27
Kimi-Audio-7B-InstructKimi Text-to-speech / audio 9.77B 8K 22.0 GB11.9 GB7.0 GB 8GB GPU Commercial OK 2025-04-25
Llama-Guard-4-12BLlama Safety / guard 12B — 26.9 GB14.5 GB8.5 GB 12GB GPU Commercial with conditions 2025-04-23
OLMo-2-0425-1BAi2 OLMo / Molmo Text LLM 1.49B 4K 4.8 GB3.3 GB2.6 GB 8GB GPU Commercial OK 2025-04-17
Phi-4-reasoning-plusPhi · reasoning Text LLM 14.7B 32K 34.4 GB19.3 GB12.0 GB 16GB GPU Commercial OK 2025-04-17
Wan2.1-FLF2V-14B-720PWan Video generation 16.4B — 36.6 GB19.7 GB11.4 GB 16GB GPU Commercial OK 2025-04-17
UI-TARS-1.5-7BByteDance Seed Vision-language 8.29B 125K 19.2 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2025-04-16
Nemotron-H-4B-Instruct-128KNVIDIA Nemotron Text LLM 4.49B 128K 10.5 GB5.8 GB3.6 GB 8GB GPU Non-commercial 2025-04-15
parakeet-tdt-0.6b-v2NVIDIA Parakeet / Canary speech Speech-to-text 600M — 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2025-04-15
granite-speech-3.3-8bIBM Granite Speech-to-text 8.65B 128K 19.5 GB10.6 GB6.3 GB 8GB GPU Commercial OK 2025-04-14
granite-3.3-2b-instructIBM Granite Text LLM 2.53B 128K 6.7 GB4.1 GB2.9 GB 8GB GPU Commercial OK 2025-04-09
granite-3.3-8b-instructIBM Granite Text LLM 8.17B 128K 19.8 GB11.4 GB7.3 GB 12GB GPU Commercial OK 2025-04-09
Kimi-VL-A3B-InstructKimi Vision-language 16.4B / 3B 128K 36.8 GB19.9 GB11.7 GB 16GB GPU Commercial OK 2025-04-09
Kimi-VL-A3B-ThinkingKimi · reasoning Vision-language 16.4B / 3B 128K 36.8 GB19.9 GB11.7 GB 16GB GPU Commercial OK 2025-04-09
Phi-4-reasoningPhi · reasoning Text LLM 14.7B 32K 34.4 GB19.3 GB12.0 GB 16GB GPU Commercial OK 2025-04-09
gemma-3-12b-it-qat-q4_0-unquantizedGemma Vision-language 12.2B — 27.3 GB14.7 GB8.6 GB 12GB GPU Commercial with conditions 2025-04-08
GLM-Z1-9B-0414GLM / CogVideoX (Z.ai) Text LLM 9.4B 32K 21.5 GB11.8 GB7.1 GB 8GB GPU Commercial OK 2025-04-08
GLM-Z1-32B-0414GLM / CogVideoX (Z.ai) Text LLM 32.6B 32K 72.7 GB39.1 GB22.7 GB RTX 5090 32GB Commercial OK 2025-04-08
Nemotron-H-56B-Base-8KNVIDIA Nemotron Text LLM 56.3B 8K 125 GB66.7 GB38.4 GB 2 × RTX 5090 (64GB) Non-commercial 2025-04-08
GLM-4-9B-0414GLM / CogVideoX (Z.ai) Text LLM 9.4B 32K 21.5 GB11.8 GB7.1 GB 8GB GPU Commercial OK 2025-04-07
GLM-4-32B-0414GLM / CogVideoX (Z.ai) Text LLM 32.6B 32K 72.7 GB39.1 GB22.7 GB RTX 5090 32GB Commercial OK 2025-04-07
Llama-3_1-Nemotron-Ultra-253B-v1NVIDIA Nemotron Text LLM 253.4B 128K 560 GB299 GB172 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2025-04-07
Llama-4-Scout-17B-16E-InstructLlama Vision-language 108.6B — 240 GB128 GB73.0 GB RTX PRO 6000 96GB Commercial with conditions 2025-04-02
Llama-4-Maverick-17B-128E-InstructLlama Vision-language 401.6B — 884 GB470 GB268 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2025-04-01
jina-reranker-m0Jina embeddings & rerankers Reranker 2.44B 32K 5.9 GB3.4 GB2.1 GB 8GB GPU Non-commercial 2025-03-27
granite-speech-3.2-8bIBM Granite Speech-to-text 8.48B 128K 19.2 GB10.4 GB6.2 GB 8GB GPU Commercial OK 2025-03-26
DeepSeek-V3-0324DeepSeek Text LLM 684.5B MoE 160K 1.51 TB801 GB458 GB 8 × H100 (640GB) Commercial OK 2025-03-24
Qwen2.5-Omni-7BQwen Omni / audio LLM 10.7B 32K 24.6 GB13.5 GB8.1 GB 12GB GPU Commercial OK 2025-03-22
Qwen2.5-VL-32B-InstructQwen Vision-language 33.5B 125K 76.2 GB41.7 GB25.0 GB RTX 5090 32GB Commercial OK 2025-03-21
Nemotron-H-4B-Base-8KNVIDIA Nemotron Text LLM 4.49B 8K 10.5 GB5.8 GB3.6 GB 8GB GPU Non-commercial 2025-03-20
Nemotron-H-8B-Base-8KNVIDIA Nemotron Text LLM 8.1B 8K 18.5 GB10.1 GB6.0 GB 8GB GPU Non-commercial 2025-03-19
LTX-Video-0.9.5LTX Video Video generation 1.92B — 4.7 GB2.7 GB1.8 GB 8GB GPU Custom licence – read it 2025-03-16
Llama-3.1-Nemotron-Nano-8B-v1NVIDIA Nemotron Text LLM 8.03B 128K 19.2 GB11.0 GB6.9 GB 8GB GPU Commercial with conditions 2025-03-16
Llama-3_3-Nemotron-Super-49B-v1NVIDIA Nemotron Text LLM 49.9B 128K 112 GB60.4 GB35.4 GB 2 × RTX 5090 (64GB) Commercial with conditions 2025-03-16
OLMo-2-0325-32B-InstructAi2 OLMo / Molmo Text LLM 32.2B 4K 73.6 GB40.3 GB24.1 GB RTX 5090 32GB Commercial OK 2025-03-12
c4ai-command-a-03-2025Cohere Command / Aya Text LLM 111.1B — 245 GB130 GB74.6 GB RTX PRO 6000 96GB Non-commercial 2025-03-11
Mistral-Small-3.1-24B-Instruct-2503Mistral / Magistral / Devstral Text LLM 24B 128K 54.7 GB29.9 GB17.9 GB RTX 4090 24GB Commercial OK 2025-03-11
canary-180m-flashNVIDIA Parakeet / Canary speech Speech-to-text 180M — 0.9 GB0.7 GB0.6 GB 8GB GPU Commercial OK 2025-03-11
gemma-3-1b-itGemma Text LLM 1B — 2.7 GB1.7 GB1.2 GB 8GB GPU Commercial with conditions 2025-03-10
canary-1b-flashNVIDIA Parakeet / Canary speech Speech-to-text 811M — 2.3 GB1.4 GB1.0 GB 8GB GPU Commercial OK 2025-03-07
QwQ-32BQwen · reasoning Text LLM 32.8B 40K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2025-03-05
shieldgemma-2-4b-itGemma Safety / guard 4.3B — 10.0 GB5.5 GB3.4 GB 8GB GPU Commercial with conditions 2025-03-04
Llama-3.1-Nemotron-8B-UltraLong-1M-InstructNVIDIA Nemotron Text LLM 8.04B 1M 19.3 GB11.0 GB6.9 GB 8GB GPU Non-commercial 2025-03-04
CogView4-6BGLM / CogVideoX (Z.ai) Image generation 6.37B — 14.5 GB7.9 GB4.7 GB 8GB GPU Commercial OK 2025-03-03
aya-vision-8bCohere Command / Aya Vision-language 8.63B — 19.5 GB10.6 GB6.3 GB 8GB GPU Non-commercial 2025-03-02
gemma-3-12b-itGemma Vision-language 12.2B — 27.3 GB14.7 GB8.6 GB 12GB GPU Commercial with conditions 2025-03-01
gemma-3-27b-itGemma Vision-language 27.4B — 60.9 GB32.6 GB18.8 GB RTX 4090 24GB Commercial with conditions 2025-03-01
Wan2.1-I2V-14B-720P-DiffusersWan Video generation 16.4B — 36.6 GB19.7 GB11.4 GB 16GB GPU Commercial OK 2025-03-01
Wan2.1-T2V-1.3B-DiffusersWan Video generation 1.42B — 3.6 GB2.2 GB1.4 GB 8GB GPU Commercial OK 2025-03-01
Wan2.1-T2V-14B-DiffusersWan Video generation 14.3B — 31.9 GB17.2 GB10.0 GB 12GB GPU Commercial OK 2025-03-01
Ling-liteAnt Ling / Ring / Ming Text LLM 16.8B / 2.75B 32K 37.9 GB20.6 GB12.2 GB 16GB GPU Commercial OK 2025-02-28
Ling-plusAnt Ling / Ring / Ming Text LLM 290B / 28.8B 16K 642 GB342 GB197 GB 4 × RTX PRO 6000 (384GB) Commercial OK 2025-02-28
c4ai-command-r7b-arabic-02-2025Cohere Command / Aya Text LLM 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Non-commercial 2025-02-27
BGE-VL-baseBGE embeddings & rerankers Embedding 150M 0K 0.8 GB0.7 GB0.6 GB 8GB GPU Commercial OK 2025-02-25
BGE-VL-largeBGE embeddings & rerankers Embedding 428M 0K 1.4 GB1.0 GB0.8 GB 8GB GPU Commercial OK 2025-02-25
Wan2.1-I2V-14B-480PWan Video generation 16.4B — 36.6 GB19.7 GB11.4 GB 16GB GPU Commercial OK 2025-02-25
Phi-4-multimodal-instructPhi Speech-to-text 5.57B 128K 12.8 GB7.0 GB4.2 GB 8GB GPU Commercial OK 2025-02-24
Moonlight-16B-A3B-InstructKimi Text LLM 16B / 3B 8K 35.9 GB19.4 GB11.4 GB 16GB GPU Commercial OK 2025-02-22
gemma-3-4b-itGemma Vision-language 4.3B — 10.0 GB5.5 GB3.4 GB 8GB GPU Commercial with conditions 2025-02-20
sv4d2.0Stable Diffusion / Stable Audio Text LLM — — ——— — Commercial with conditions 2025-02-20
Phi-4-mini-instructPhi Text LLM 3.84B 128K 10.0 GB6.1 GB4.1 GB 8GB GPU Commercial OK 2025-02-19
granite-3.2-2b-instructIBM Granite Text LLM 2.53B 128K 6.7 GB4.1 GB2.9 GB 8GB GPU Commercial OK 2025-02-17
granite-3.2-8b-instructIBM Granite Text LLM 8.17B 128K 19.8 GB11.4 GB7.3 GB 12GB GPU Commercial OK 2025-02-17
granite-embedding-30m-sparseIBM Granite Embedding 30M 1K 0.6 GB0.5 GB0.5 GB 8GB GPU Commercial OK 2025-02-17
granite-vision-3.2-2bIBM Granite Vision-language 2.98B 128K 7.7 GB4.6 GB3.2 GB 8GB GPU Commercial OK 2025-02-17
Step-Audio-ChatStepFun Step Omni / audio LLM 131.9B — 326 GB190 GB124 GB H200 141GB Commercial OK 2025-02-13
SmolVLM2-2.2B-InstructSmolLM / SmolVLM Vision-language 2.25B 8K 5.4 GB3.1 GB2.0 GB 8GB GPU Commercial OK 2025-02-08
nemocurator-fineweb-nemotron-4-edu-classifierNVIDIA Nemotron Text LLM 109M 1K 1.0 GB0.9 GB0.9 GB 8GB GPU Custom licence – read it 2025-02-07
granite-guardian-3.2-3b-a800mIBM Granite Safety / guard 3.3B MoE 128K 8.3 GB4.9 GB3.2 GB 8GB GPU Commercial OK 2025-02-03
Mistral-Small-24B-Instruct-2501Mistral / Magistral / Devstral Text LLM 23.6B 32K 53.7 GB29.4 GB17.6 GB RTX 4090 24GB Commercial OK 2025-01-28
OLMoE-1B-7B-0125-InstructAi2 OLMo / Molmo Text LLM 6.92B MoE 4K 16.8 GB9.7 GB6.2 GB 8GB GPU Commercial OK 2025-01-27
Qwen2.5-VL-72B-InstructQwen Vision-language 73.4B 125K 165 GB89.0 GB52.1 GB 2 × RTX 5090 (64GB) Commercial with conditions 2025-01-27
Qwen2.5-VL-3B-InstructQwen Vision-language 3.75B 125K 9.1 GB5.2 GB3.3 GB 8GB GPU Non-commercial 2025-01-26
Qwen2.5-VL-7B-InstructQwen Vision-language 8.29B 125K 19.2 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2025-01-26
granite-guardian-3.2-5bIBM Granite Safety / guard 5.78B 128K 14.2 GB8.2 GB5.3 GB 8GB GPU Commercial OK 2025-01-23
Qwen2.5-7B-Instruct-1MQwen Text LLM 7.62B 1M 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2025-01-23
Qwen2.5-14B-Instruct-1MQwen Text LLM 14.8B 1M 34.6 GB19.4 GB12.0 GB 16GB GPU Commercial OK 2025-01-23
DeepSeek-R1DeepSeek · reasoning Text LLM 671B / 37B 160K 1.48 TB785 GB449 GB 8 × H100 (640GB) Commercial OK 2025-01-20
DeepSeek-R1-Distill-Llama-8BDeepSeek · reasoning Text LLM 8.03B 128K 19.2 GB11.0 GB6.9 GB 8GB GPU Commercial OK 2025-01-20
DeepSeek-R1-Distill-Llama-70BDeepSeek · reasoning Text LLM 70.6B 128K 158 GB85.6 GB50.2 GB 2 × RTX 5090 (64GB) Commercial OK 2025-01-20
DeepSeek-R1-Distill-Qwen-1.5BDeepSeek · reasoning Text LLM 1.78B 128K 4.6 GB2.8 GB1.9 GB 8GB GPU Commercial OK 2025-01-20
DeepSeek-R1-Distill-Qwen-7BDeepSeek · reasoning Text LLM 7.62B 128K 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2025-01-20
DeepSeek-R1-Distill-Qwen-14BDeepSeek · reasoning Text LLM 14.8B 128K 34.6 GB19.4 GB12.0 GB 16GB GPU Commercial OK 2025-01-20
DeepSeek-R1-Distill-Qwen-32BDeepSeek · reasoning Text LLM 32.8B 128K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2025-01-20
DeepSeek-R1-ZeroDeepSeek · reasoning Text LLM 671B / 37B 160K 1.48 TB785 GB449 GB 8 × H100 (640GB) Commercial OK 2025-01-20
MiniCPM-Embedding-LightMiniCPM / VoxCPM Embedding 434M 4K 1.5 GB1.0 GB0.8 GB 8GB GPU Licence not stated 2025-01-17
olmOCR-7B-0225-previewAi2 OLMo / Molmo Vision-language 8.29B 32K 19.2 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2025-01-15
internlm3-8b-instructInternLM / Intern-S Text LLM 8.8B 32K 20.3 GB11.2 GB6.8 GB 8GB GPU Commercial OK 2025-01-13
MiniCPM-o-2_6MiniCPM / VoxCPM Omni / audio LLM 8.68B 32K 20.1 GB11.1 GB6.8 GB 8GB GPU Commercial OK 2025-01-12
MiniMax-VL-01MiniMax Vision-language 456.4B MoE 8K 1.00 TB534 GB305 GB 4 × RTX PRO 6000 (384GB) Licence not stated 2025-01-12
Llama-3.1-Tulu-3-405BAi2 OLMo / Molmo Text LLM 405.9B 128K 898 GB479 GB275 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2025-01-09
DeepSeek-V3DeepSeek Text LLM 671B / 37B 160K 1.48 TB785 GB449 GB 8 × H100 (640GB) Licence not stated 2024-12-25
QVQ-72B-PreviewQwen Vision-language 73.4B 125K 165 GB89.0 GB52.1 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-12-24
granite-guardian-3.1-2bIBM Granite Safety / guard 2.53B 128K 6.7 GB4.1 GB2.9 GB 8GB GPU Commercial OK 2024-12-17
granite-guardian-3.1-8bIBM Granite Safety / guard 8.17B 128K 19.8 GB11.4 GB7.3 GB 12GB GPU Commercial OK 2024-12-17
deepseek-vl2DeepSeek Vision-language 27.5B — 61.0 GB32.6 GB18.8 GB RTX 4090 24GB Commercial with conditions 2024-12-13
deepseek-vl2-smallDeepSeek Vision-language 16.1B / 2.8B 4K 37.8 GB21.2 GB13.1 GB 16GB GPU Commercial with conditions 2024-12-13
deepseek-vl2-tinyDeepSeek Vision-language 3.37B / 1B 4K 8.4 GB4.9 GB3.2 GB 8GB GPU Commercial with conditions 2024-12-13
c4ai-command-r7b-12-2024Cohere Command / Aya Text LLM 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Non-commercial 2024-12-11
phi-4Phi Text LLM 14.7B 16K 34.4 GB19.3 GB12.0 GB 16GB GPU Commercial OK 2024-12-11
DeepSeek-V2.5-1210DeepSeek Text LLM 235.7B MoE 160K 520 GB277 GB158 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2024-12-10
granite-3.1-2b-instructIBM Granite Text LLM 2.53B 128K 6.7 GB4.1 GB2.9 GB 8GB GPU Commercial OK 2024-12-06
granite-3.1-8b-instructIBM Granite Text LLM 8.17B 128K 19.8 GB11.4 GB7.3 GB 12GB GPU Commercial OK 2024-12-06
granite-3.1-1b-a400m-instructIBM Granite Text LLM 1.33B MoE 128K 3.8 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-12-06
granite-3.1-3b-a800m-instructIBM Granite Text LLM 3.3B MoE 128K 8.3 GB4.9 GB3.2 GB 8GB GPU Commercial OK 2024-12-06
granite-embedding-30m-englishIBM Granite Embedding 30M 1K 0.6 GB0.5 GB0.5 GB 8GB GPU Commercial OK 2024-12-04
granite-embedding-125m-englishIBM Granite Embedding 125M 1K 0.8 GB0.6 GB0.6 GB 8GB GPU Commercial OK 2024-12-04
granite-embedding-107m-multilingualIBM Granite Embedding 107M 1K 0.7 GB0.6 GB0.6 GB 8GB GPU Commercial OK 2024-12-04
granite-embedding-278m-multilingualIBM Granite Embedding 278M 1K 1.1 GB0.8 GB0.7 GB 8GB GPU Commercial OK 2024-12-04
HunyuanVideoTencent Hunyuan / Hy Video generation — — ——— — Commercial with conditions 2024-12-01
QwQ-32B-PreviewQwen · reasoning Text LLM 32.8B 32K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2024-11-27
Llama-3.3-70B-InstructLlama Text LLM 70.6B 128K 158 GB85.6 GB50.2 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-11-26
GOT-OCR-2.0-hfStepFun Step Vision-language 561M — 2.5 GB2.0 GB1.7 GB 8GB GPU Commercial OK 2024-11-22
paligemma2-3b-pt-224Gemma Vision-language 3.03B — 7.2 GB4.0 GB2.5 GB 8GB GPU Commercial with conditions 2024-11-21
paligemma2-3b-pt-448Gemma Vision-language 3.03B — 7.2 GB4.0 GB2.5 GB 8GB GPU Commercial with conditions 2024-11-21
paligemma2-3b-ft-docci-448Gemma Vision-language 3.03B — 7.2 GB4.0 GB2.5 GB 8GB GPU Commercial with conditions 2024-11-21
paligemma2-10b-ft-docci-448Gemma Vision-language 9.66B — 21.8 GB11.8 GB6.9 GB 8GB GPU Commercial with conditions 2024-11-21
paligemma2-3b-mix-224Gemma Vision-language 3.03B — 7.2 GB4.0 GB2.5 GB 8GB GPU Commercial with conditions 2024-11-21
paligemma2-3b-mix-448Gemma Vision-language 3.03B — 7.2 GB4.0 GB2.5 GB 8GB GPU Commercial with conditions 2024-11-21
FLUX.1-Canny-devFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Non-commercial 2024-11-20
FLUX.1-Depth-devFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Non-commercial 2024-11-20
FLUX.1-Fill-devFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Non-commercial 2024-11-20
FLUX.1-Redux-devFLUX Image generation — — ——— — Non-commercial 2024-11-20
glm-edge-1.5b-chatGLM / CogVideoX (Z.ai) Text LLM 1.59B 8K 4.5 GB2.8 GB2.0 GB 8GB GPU Custom licence – read it 2024-11-20
glm-edge-4b-chatGLM / CogVideoX (Z.ai) Text LLM 4.33B 8K 11.0 GB6.6 GB4.4 GB 8GB GPU Custom licence – read it 2024-11-20
Llama-3.1-Tulu-3-8BAi2 OLMo / Molmo Text LLM 8.03B 128K 19.2 GB11.0 GB6.9 GB 8GB GPU Commercial with conditions 2024-11-20
OLMo-2-1124-13BAi2 OLMo / Molmo Text LLM 13.7B 4K 37.4 GB23.2 GB16.4 GB RTX 4090 24GB Commercial OK 2024-11-19
Mistral-Large-Instruct-2411Mistral / Magistral / Devstral Text LLM 122.6B 128K 273 GB147 GB85.2 GB RTX PRO 6000 96GB Non-commercial 2024-11-14
Qwen2.5-Coder-0.5B-InstructQwen · coding Text LLM 494M 32K 1.7 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2024-11-06
Qwen2.5-Coder-3B-InstructQwen · coding Text LLM 3.09B 32K 7.6 GB4.4 GB2.9 GB 8GB GPU Non-commercial 2024-11-06
Qwen2.5-Coder-14B-InstructQwen · coding Text LLM 14.8B 32K 34.6 GB19.4 GB12.0 GB 16GB GPU Commercial OK 2024-11-06
Qwen2.5-Coder-32B-InstructQwen · coding Text LLM 32.8B 32K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2024-11-06
CogVideoX1.5-5B-I2VGLM / CogVideoX (Z.ai) Video generation 5.57B — 12.8 GB7.0 GB4.2 GB 8GB GPU Custom licence – read it 2024-11-02
LTX-VideoLTX Video Video generation 1.92B — 4.7 GB2.7 GB1.8 GB 8GB GPU Custom licence – read it 2024-10-31
SmolLM2-135MSmolLM / SmolVLM Text LLM 135M 8K 1.0 GB0.8 GB0.8 GB 8GB GPU Commercial OK 2024-10-31
SmolLM2-360MSmolLM / SmolVLM Text LLM 362M 8K 1.6 GB1.3 GB1.1 GB 8GB GPU Commercial OK 2024-10-31
SmolLM2-1.7BSmolLM / SmolVLM Text LLM 1.71B 8K 5.9 GB4.1 GB3.3 GB 8GB GPU Commercial OK 2024-10-30
OLMo-2-1124-7BAi2 OLMo / Molmo Text LLM 7.3B 4K 20.8 GB13.3 GB9.7 GB 12GB GPU Commercial OK 2024-10-29
stable-diffusion-3.5-mediumStable Diffusion / Stable Audio Image generation 2.47B — 5.9 GB3.4 GB2.1 GB 8GB GPU Commercial with conditions 2024-10-29
aya-expanse-8bCohere Command / Aya Text LLM 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Non-commercial 2024-10-23
aya-expanse-32bCohere Command / Aya Text LLM 32.3B — 71.6 GB38.2 GB22.0 GB RTX 5090 32GB Non-commercial 2024-10-23
Nemotron-4-Mini-Hindi-4B-InstructNVIDIA Nemotron Text LLM 4.19B 4K 10.8 GB6.5 GB4.4 GB 8GB GPU Commercial with conditions 2024-10-22
stable-diffusion-3.5-largeStable Diffusion / Stable Audio Image generation 8.15B — 18.4 GB10.0 GB5.9 GB 8GB GPU Commercial with conditions 2024-10-22
stable-diffusion-3.5-large-turboStable Diffusion / Stable Audio Image generation 8.15B — 18.4 GB10.0 GB5.9 GB 8GB GPU Commercial with conditions 2024-10-22
granite-guardian-3.0-2bIBM Granite Safety / guard 2.53B 8K 6.7 GB4.1 GB2.9 GB 8GB GPU Commercial OK 2024-10-15
Ministral-8B-Instruct-2410Mistral / Magistral / Devstral Text LLM 8.02B 32K 19.4 GB11.1 GB7.1 GB 8GB GPU Non-commercial 2024-10-15
Llama-3.1-Nemotron-70B-Instruct-HFNVIDIA Nemotron Text LLM 70.6B 128K 158 GB85.6 GB50.2 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-10-12
granite-3.0-1b-a400m-instructIBM Granite Text LLM 1.33B / 400M 4K 3.8 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-10-03
granite-3.0-3b-a800m-instructIBM Granite Text LLM 3.37B / 800M 4K 8.5 GB5.0 GB3.3 GB 8GB GPU Commercial OK 2024-10-03
granite-3.0-2b-instructIBM Granite Text LLM 2.63B 4K 7.0 GB4.2 GB2.9 GB 8GB GPU Commercial OK 2024-10-02
granite-3.0-8b-instructIBM Granite Text LLM 8.17B 4K 19.8 GB11.4 GB7.3 GB 12GB GPU Commercial OK 2024-10-02
whisper-large-v3-turboWhisper Speech-to-text 809M — 2.3 GB1.4 GB1.0 GB 8GB GPU Commercial OK 2024-10-01
gemma-2-2b-jpn-itGemma Text LLM 2.61B — 6.3 GB3.6 GB2.2 GB 8GB GPU Commercial with conditions 2024-09-25
Molmo-72B-0924Ai2 OLMo / Molmo Vision-language 73.3B 4K 165 GB88.9 GB52.1 GB 2 × RTX 5090 (64GB) Commercial OK 2024-09-25
Molmo-7B-D-0924Ai2 OLMo / Molmo Vision-language 8.02B 4K 18.6 GB10.3 GB6.3 GB 8GB GPU Commercial OK 2024-09-25
Molmo-7B-O-0924Ai2 OLMo / Molmo Vision-language 7.67B 4K 21.7 GB13.7 GB9.9 GB 12GB GPU Commercial OK 2024-09-25
MolmoE-1B-0924Ai2 OLMo / Molmo Vision-language 7.2B / 1.5B 32K 16.3 GB8.9 GB5.3 GB 8GB GPU Commercial OK 2024-09-24
Llama-3_1-Nemotron-51B-InstructNVIDIA Nemotron Text LLM 51.5B 128K 115 GB62.3 GB36.4 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-09-22
Llama-Guard-3-1BLlama Safety / guard 1.5B — 3.8 GB2.3 GB1.5 GB 8GB GPU Commercial with conditions 2024-09-20
Llama-Guard-3-11B-VisionLlama Safety / guard 10.7B — 24.0 GB13.0 GB7.6 GB 12GB GPU Commercial with conditions 2024-09-20
Qwen2.5-Math-7B-InstructQwen Text LLM 7.62B 4K 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2024-09-19
Llama-3.2-1B-InstructLlama Text LLM 1.24B 128K 3.5 GB2.2 GB1.6 GB 8GB GPU Commercial with conditions 2024-09-18
Llama-3.2-3B-InstructLlama Text LLM 3.21B 128K 8.5 GB5.2 GB3.6 GB 8GB GPU Commercial with conditions 2024-09-18
Llama-3.2-11B-Vision-InstructLlama Vision-language 10.7B — 24.0 GB13.0 GB7.6 GB 12GB GPU Commercial with conditions 2024-09-18
Qwen2.5-Coder-1.5B-InstructQwen · coding Text LLM 1.54B 32K 4.1 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-09-18
Mistral-Small-Instruct-2409Mistral / Magistral / Devstral Text LLM 22.2B 32K 51.3 GB28.4 GB17.2 GB RTX 4090 24GB Non-commercial 2024-09-17
parakeet-tdt_ctc-110mNVIDIA Parakeet / Canary speech Speech-to-text 110M — 0.7 GB0.6 GB0.6 GB 8GB GPU Commercial OK 2024-09-17
Qwen2-VL-72B-InstructQwen Vision-language 73.4B 32K 165 GB89.0 GB52.1 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-09-17
Qwen2.5-1.5B-InstructQwen Text LLM 1.54B 32K 4.1 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-09-17
Qwen2.5-3B-InstructQwen Text LLM 3.09B 32K 7.6 GB4.4 GB2.9 GB 8GB GPU Non-commercial 2024-09-17
Qwen2.5-32B-InstructQwen Text LLM 32.8B 32K 74.7 GB40.9 GB24.5 GB RTX 5090 32GB Commercial OK 2024-09-17
Qwen2.5-Coder-7B-InstructQwen · coding Text LLM 7.62B 32K 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2024-09-17
CogVideoX-5b-I2VGLM / CogVideoX (Z.ai) Video generation 5.63B — 12.9 GB7.1 GB4.3 GB 8GB GPU Custom licence – read it 2024-09-16
Qwen2.5-0.5B-InstructQwen Text LLM 494M 32K 1.7 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2024-09-16
Qwen2.5-7B-InstructQwen Text LLM 7.62B 32K 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2024-09-16
Qwen2.5-14B-InstructQwen Text LLM 14.8B 32K 34.6 GB19.4 GB12.0 GB 16GB GPU Commercial OK 2024-09-16
Qwen2.5-72B-InstructQwen Text LLM 72.7B 32K 163 GB88.2 GB51.7 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-09-16
Qwen2.5-Math-1.5BQwen Text LLM 1.54B 4K 4.1 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-09-16
Qwen2.5-Math-72B-InstructQwen Text LLM 72.7B 4K 163 GB88.2 GB51.7 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-09-16
GOT-OCR2_0StepFun Step Vision-language 716M 32K 2.9 GB2.1 GB1.8 GB 8GB GPU Commercial OK 2024-09-12
Pixtral-12B-2409Mistral / Magistral / Devstral Text LLM 12B 128K 28.2 GB15.9 GB9.8 GB 12GB GPU Commercial OK 2024-09-11
Nemotron-Mini-4B-InstructNVIDIA Nemotron Text LLM 4B 4K 10.4 GB6.2 GB4.2 GB 8GB GPU Commercial with conditions 2024-09-10
DeepSeek-V2.5DeepSeek Text LLM 235.7B MoE 160K 520 GB277 GB158 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2024-09-05
DeepSeek-Coder-V2-Instruct-0724DeepSeek · coding Text LLM 236B / 21B 160K 520 GB277 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2024-09-05
granite-guardian-hap-38mIBM Granite Safety / guard 38M 1K 0.7 GB0.6 GB0.6 GB 8GB GPU Commercial OK 2024-09-05
granite-guardian-hap-125mIBM Granite Safety / guard 125M 1K 1.1 GB0.9 GB0.9 GB 8GB GPU Commercial OK 2024-09-05
jina-embeddings-v3Jina embeddings & rerankers Embedding 572M 8K 1.8 GB1.2 GB0.9 GB 8GB GPU Non-commercial 2024-09-05
MiniCPM-EmbeddingMiniCPM / VoxCPM Embedding 2.73B 1K 6.5 GB3.7 GB2.3 GB 8GB GPU Licence not stated 2024-09-04
MiniCPM3-4BMiniCPM / VoxCPM Text LLM 4B 32K 9.6 GB5.5 GB3.5 GB 8GB GPU Commercial OK 2024-09-03
Qwen2-VL-2B-InstructQwen Vision-language 2.21B 32K 5.6 GB3.3 GB2.2 GB 8GB GPU Commercial OK 2024-08-28
Qwen2-VL-7B-InstructQwen Vision-language 8.29B 32K 19.2 GB10.7 GB6.5 GB 8GB GPU Commercial OK 2024-08-28
Yi-Coder-1.5B-ChatYi (01.AI) · coding Text LLM 1.48B 128K 5.4 GB3.8 GB3.1 GB 8GB GPU Commercial OK 2024-08-21
Yi-Coder-9B-ChatYi (01.AI) · coding Text LLM 8.83B 128K 20.7 GB11.6 GB7.2 GB 8GB GPU Commercial OK 2024-08-21
CogVideoX-5bGLM / CogVideoX (Z.ai) Video generation 5.57B — 12.8 GB7.0 GB4.2 GB 8GB GPU Custom licence – read it 2024-08-17
Phi-3.5-MoE-instructPhi Text LLM 41.9B / 6.6B 128K 93.8 GB50.5 GB29.5 GB 2 × RTX 5090 (64GB) Commercial OK 2024-08-17
jina-colbert-v2Jina embeddings & rerankers Embedding 559M 8K 1.7 GB1.2 GB0.9 GB 8GB GPU Non-commercial 2024-08-16
Phi-3.5-mini-instructPhi Text LLM 3.82B 128K 12.1 GB8.2 GB6.3 GB 8GB GPU Commercial OK 2024-08-16
Phi-3.5-vision-instructPhi Vision-language 4.15B 128K 12.8 GB8.6 GB6.5 GB 8GB GPU Commercial OK 2024-08-16
Qwen2-Math-1.5BQwen Text LLM 1.54B 4K 4.1 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-08-08
CogVideoX-2bGLM / CogVideoX (Z.ai) Video generation 1.69B — 4.2 GB2.5 GB1.6 GB 8GB GPU Commercial OK 2024-08-05
MiniCPM-V-2_6MiniCPM / VoxCPM Vision-language 8.1B — 18.3 GB10.0 GB5.9 GB 8GB GPU Licence not stated 2024-08-04
FLUX.1-devFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Non-commercial 2024-07-31
FLUX.1-schnellFLUX Image generation 11.9B — 26.7 GB14.4 GB8.4 GB 12GB GPU Commercial OK 2024-07-31
Qwen2-Audio-7B-InstructQwen Omni / audio LLM 8.4B — 19.0 GB10.3 GB6.1 GB 8GB GPU Commercial OK 2024-07-31
internlm2_5-1_8b-chatInternLM / Intern-S Text LLM 1.89B 32K 5.5 GB3.5 GB2.6 GB 8GB GPU Custom licence – read it 2024-07-30
bge-en-iclBGE embeddings & rerankers Embedding 7.11B 32K 16.1 GB8.8 GB5.2 GB 8GB GPU Commercial OK 2024-07-25
bge-multilingual-gemma2BGE embeddings & rerankers Embedding 9.24B 8K 20.8 GB11.3 GB6.7 GB 8GB GPU Commercial with conditions 2024-07-25
bge-reranker-v2.5-gemma2-lightweightBGE embeddings & rerankers Reranker 9.24B 8K 20.8 GB11.3 GB6.7 GB 8GB GPU Commercial with conditions 2024-07-25
Mistral-Large-Instruct-2407Mistral / Magistral / Devstral Text LLM 122.6B — 270 GB144 GB82.3 GB RTX PRO 6000 96GB Non-commercial 2024-07-24
Llama-Guard-3-8BLlama Safety / guard 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Commercial with conditions 2024-07-22
Prompt-Guard-86MLlama Safety / guard 279M — 1.1 GB0.8 GB0.7 GB 8GB GPU Commercial with conditions 2024-07-21
OLMoE-1B-7B-0924Ai2 OLMo / Molmo Text LLM 6.92B MoE 4K 16.8 GB9.7 GB6.2 GB 8GB GPU Commercial OK 2024-07-20
DeepSeek-V2-Chat-0628DeepSeek Text LLM 235.7B MoE 160K 520 GB277 GB158 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2024-07-18
Llama-3.1-8B-InstructLlama Text LLM 8.03B 128K 19.2 GB11.0 GB6.9 GB 8GB GPU Commercial with conditions 2024-07-18
Mistral-Nemo-Instruct-2407Mistral / Magistral / Devstral Text LLM 12.2B 128K 28.8 GB16.2 GB10.0 GB 12GB GPU Commercial OK 2024-07-17
gemma-2-2b-itGemma Text LLM 2.61B — 6.3 GB3.6 GB2.2 GB 8GB GPU Commercial with conditions 2024-07-16
shieldgemma-9bGemma Safety / guard 9.24B — 20.8 GB11.3 GB6.7 GB 8GB GPU Commercial with conditions 2024-07-16
Llama-3.1-70B-InstructLlama Text LLM 70.6B 128K 158 GB85.6 GB50.2 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-07-16
Llama-3.1-405BLlama Text LLM 405.9B 128K 898 GB479 GB275 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2024-07-16
SmolLM-135MSmolLM / SmolVLM Text LLM 135M 2K 1.0 GB0.8 GB0.8 GB 8GB GPU Commercial OK 2024-07-14
SmolLM-360MSmolLM / SmolVLM Text LLM 362M 2K 1.6 GB1.3 GB1.1 GB 8GB GPU Commercial OK 2024-07-14
SmolLM-1.7BSmolLM / SmolVLM Text LLM 1.71B 2K 5.9 GB4.1 GB3.3 GB 8GB GPU Commercial OK 2024-07-14
granite-8b-code-instruct-128kIBM Granite · coding Text LLM 8.05B 125K 19.4 GB11.1 GB7.1 GB 8GB GPU Commercial OK 2024-07-12
OLMo-7B-0724-hfAi2 OLMo / Molmo Text LLM 6.89B 4K 19.9 GB12.8 GB9.4 GB 12GB GPU Commercial OK 2024-07-12
codegeex4-all-9bGLM / CogVideoX (Z.ai) · coding Text LLM 9.4B 128K 26.6 GB16.9 GB12.1 GB 16GB GPU Custom licence – read it 2024-07-05
SenseVoiceSmallFunAudioLLM speech Speech-to-text — — ——— — Custom licence – read it 2024-07-03
internlm2_5-7b-chatInternLM / Intern-S Text LLM 7.74B 32K 18.6 GB10.6 GB6.7 GB 8GB GPU Custom licence – read it 2024-06-27
gemma-2-9b-itGemma Text LLM 9.24B — 20.8 GB11.3 GB6.7 GB 8GB GPU Commercial with conditions 2024-06-24
gemma-2-27b-itGemma Text LLM 27.2B — 60.4 GB32.3 GB18.7 GB RTX 4090 24GB Commercial with conditions 2024-06-24
jina-reranker-v2-base-multilingualJina embeddings & rerankers Reranker 278M 1K 1.1 GB0.8 GB0.7 GB 8GB GPU Non-commercial 2024-06-19
OLMo-1B-0724-hfAi2 OLMo / Molmo Text LLM 1.28B 4K 4.4 GB3.1 GB2.4 GB 8GB GPU Commercial OK 2024-06-15
DeepSeek-Coder-V2-InstructDeepSeek · coding Text LLM 236B / 21B 160K 520 GB277 GB159 GB 4 × RTX PRO 6000 (384GB) Commercial with conditions 2024-06-14
DeepSeek-Coder-V2-Lite-InstructDeepSeek · coding Text LLM 16B / 2.4B 160K 36.0 GB19.5 GB11.4 GB 16GB GPU Commercial with conditions 2024-06-14
stable-diffusion-3-medium-diffusersStable Diffusion / Stable Audio Image generation 2.08B — 5.1 GB2.9 GB1.9 GB 8GB GPU Non-commercial 2024-06-12
glm-4-9b-chatGLM / CogVideoX (Z.ai) Text LLM 9.4B 128K 26.6 GB16.9 GB12.1 GB 16GB GPU Custom licence – read it 2024-06-04
glm-4-9b-chat-1mGLM / CogVideoX (Z.ai) Text LLM 9.48B 1M 26.7 GB17.0 GB12.2 GB 16GB GPU Custom licence – read it 2024-06-04
glm-4v-9bGLM / CogVideoX (Z.ai) Text LLM 13.9B 8K 36.5 GB22.1 GB15.1 GB RTX 4090 24GB Custom licence – read it 2024-06-04
Qwen2-7BQwen Text LLM 7.62B 128K 17.7 GB9.9 GB6.0 GB 8GB GPU Commercial OK 2024-06-04
Qwen2-57B-A14B-InstructQwen Text LLM 57.4B / 14B 32K 127 GB68.1 GB39.3 GB 2 × RTX 5090 (64GB) Commercial OK 2024-06-04
Qwen2-0.5B-InstructQwen Text LLM 494M 32K 1.7 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2024-06-03
Qwen2-1.5B-InstructQwen Text LLM 1.54B 32K 4.1 GB2.5 GB1.8 GB 8GB GPU Commercial OK 2024-06-03
Codestral-22B-v0.1Mistral / Magistral / Devstral · coding Text LLM 22.2B 32K 51.3 GB28.4 GB17.2 GB RTX 4090 24GB Non-commercial 2024-05-29
internlm2-math-plus-7bInternLM / Intern-S Text LLM 7.74B 8K 18.6 GB10.6 GB6.7 GB 8GB GPU Custom licence – read it 2024-05-24
internlm2-math-plus-mixtral8x22bInternLM / Intern-S Text LLM 140.6B MoE 64K 312 GB167 GB96.2 GB H200 141GB Custom licence – read it 2024-05-24
stable-audio-open-1.0Stable Diffusion / Stable Audio Text-to-speech / audio 1.21B — 3.2 GB1.9 GB1.3 GB 8GB GPU Commercial with conditions 2024-05-24
Mistral-7B-Instruct-v0.3Mistral / Magistral / Devstral Text LLM 7.25B 32K 17.5 GB10.0 GB6.4 GB 8GB GPU Commercial OK 2024-05-22
Qwen2-72BQwen Text LLM 72.7B 128K 163 GB88.2 GB51.7 GB 2 × RTX 5090 (64GB) Commercial with conditions 2024-05-22
aya-23-8BCohere Command / Aya Text LLM 8.03B — 18.2 GB9.9 GB5.9 GB 8GB GPU Non-commercial 2024-05-19
MiniCPM-Llama3-V-2_5MiniCPM / VoxCPM Vision-language 8.54B 8K 20.4 GB11.5 GB7.3 GB 12GB GPU Licence not stated 2024-05-19
Phi-3-vision-128k-instructPhi Vision-language 4.15B 128K 12.8 GB8.6 GB6.5 GB 8GB GPU Commercial OK 2024-05-19
cogvlm2-llama3-chat-19BGLM / CogVideoX (Z.ai) Vision-language 19.5B 8K 47.7 GB27.6 GB17.8 GB RTX 4090 24GB Custom licence – read it 2024-05-16
DeepSeek-V2-LiteDeepSeek Text LLM 16B / 2.4B 160K 36.0 GB19.5 GB11.4 GB 16GB GPU Commercial with conditions 2024-05-15
Yi-1.5-9B-Chat-16KYi (01.AI) Text LLM 8.83B 16K 20.7 GB11.6 GB7.2 GB 8GB GPU Commercial OK 2024-05-15
Yi-1.5-34B-Chat-16KYi (01.AI) Text LLM 34.4B 16K 78.2 GB42.7 GB25.4 GB RTX 5090 32GB Commercial OK 2024-05-15
Yi-1.5-9B-32KYi (01.AI) Text LLM 8.83B 32K 20.7 GB11.6 GB7.2 GB 8GB GPU Commercial OK 2024-05-15
Yi-1.5-34B-32KYi (01.AI) Text LLM 34.4B 32K 78.2 GB42.7 GB25.4 GB RTX 5090 32GB Commercial OK 2024-05-15
parakeet-tdt_ctc-0.6b-jaNVIDIA Parakeet / Canary speech Speech-to-text 600M — 1.8 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2024-05-13
Yi-1.5-6BYi (01.AI) Text LLM 6.06B 4K 14.4 GB8.1 GB5.1 GB 8GB GPU Commercial OK 2024-05-11
Yi-1.5-9B-ChatYi (01.AI) Text LLM 8.83B 4K 20.7 GB11.6 GB7.2 GB 8GB GPU Commercial OK 2024-05-10
Yi-1.5-34B-ChatYi (01.AI) Text LLM 34.4B 4K 78.2 GB42.7 GB25.4 GB RTX 5090 32GB Commercial OK 2024-05-10
parakeet-tdt_ctc-1.1bNVIDIA Parakeet / Canary speech Speech-to-text 1.1B — 2.9 GB1.8 GB1.2 GB 8GB GPU Commercial OK 2024-05-07
Phi-3-medium-128k-instructPhi Text LLM 14B 128K 32.9 GB18.5 GB11.5 GB 16GB GPU Commercial OK 2024-05-07
Phi-3-medium-4k-instructPhi Text LLM 14B 4K 32.9 GB18.5 GB11.5 GB 16GB GPU Commercial OK 2024-05-07
Phi-3-small-128k-instructPhi Text LLM 7.39B 128K 17.8 GB10.2 GB6.5 GB 8GB GPU Commercial OK 2024-05-07
Phi-3-small-8k-instructPhi Text LLM 7.39B 8K 17.8 GB10.2 GB6.5 GB 8GB GPU Commercial OK 2024-05-07
Mixtral-8x22B-Instruct-v0.1Mistral / Magistral / Devstral Text LLM 140.6B MoE 64K 312 GB167 GB96.2 GB H200 141GB Commercial OK 2024-04-16
bge-reranker-v2-m3BGE embeddings & rerankers Reranker 568M 8K 1.7 GB1.2 GB0.9 GB 8GB GPU Commercial OK 2024-03-15
bge-m3BGE embeddings & rerankers Embedding — 8K ——— — Commercial OK 2024-01-27
Mixtral-8x7B-Instruct-v0.1Mistral / Magistral / Devstral Text LLM 46.7B MoE 32K 104 GB56.2 GB32.7 GB 2 × RTX 5090 (64GB) Commercial OK 2023-12-10
whisper-large-v3Whisper Speech-to-text 1.54B — 3.9 GB2.3 GB1.5 GB 8GB GPU Commercial OK 2023-11-07

“—” = not computed: no parameter count in official metadata, or image / video / speech models whose memory depends on resolution and pipeline (see each model page). Mac columns are on model pages.

04How we estimate

Weights = parameters × bits ÷ 8: BF16 16 bit, FP8 / Q8_0 8.5 bit, Q4_K_M 4.85 bit (llama.cpp’s average).

KV cache = attention layers × 2 × KV heads × head dim × tokens × 2 bytes (FP16). MLA models (DeepSeek, Kimi, GLM-5) cache one latent per layer; sliding-window layers are capped at the window; linear-attention / Mamba layers add no KV.

Overhead = 10% of weights + 0.5GB. Mac: the GPU can use about 2/3 of unified memory up to 36GB, 3/4 above. DGX Spark: ~90%.

Parameters: from the model card when it states them (always for MoE flagships), otherwise counted from the official safetensors files or the official model name.

Not estimated: new compressed-attention designs (e.g. DeepSeek-V4, MiniMax-M3 sparse attention) – KV is left blank and only weights are counted; gated repos whose config needs a login.

Try your own context length and concurrency in the calculator →

FAQ

Which open models fit on a 24GB GPU such as the RTX 4090?

By our estimate (weights + 10% overhead + 8K-token KV cache), 382 of the 597 text / vision models here fit in 24GB at 4-bit (Q4_K_M), and 341 at FP8. Use the “Fits on” filter below to list them.

How is the VRAM number calculated?

Weights = parameters × bits per weight ÷ 8 (BF16 16 bit, FP8 / Q8_0 8.5 bit, Q4_K_M 4.85 bit). We add 10% of the weights + 0.5GB runtime overhead and the KV cache for one 8K-token request, computed from each model’s config.json (layers, KV heads, head dim; MLA, sliding-window and hybrid linear-attention layers are handled). A model fits when the total is ≤ 90% of the memory. These are estimates – engines such as vLLM pre-allocate extra memory.

For MoE models, do I size by total or active parameters?

By total parameters: every expert has to be in memory. Active parameters only decide speed (how many bytes are read per token).

Where does the model list come from?

Only official publisher organisations on Hugging Face (799 models, 1724 official repos including FP8 / GGUF / MLX builds), plus the official Ollama library for Ollama tags. We refresh it weekly and add new releases.

Data updated: Self-host open models(UTC+8)