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.

// Open-model VRAM table · family

Qwen: local deployment & VRAM

112 open-weight models from Alibaba Qwen (345 official repos incl. quantised builds). At 4-bit they range from 0.9 GB (Qwen2.5-Coder-0.5B-Instruct) to 1.60 TB (Qwen3.8-2.4T-A95B).

Filter by my hardware → VRAM calculator

official orgs: Qwenlatest 2026-10-09memory = estimate

01Text LLM (46)

ModelParamsContextBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3.8-2.4T-A95B2.4T / 95B 256K 5.28 TB1.60 TB 8 × B300 (2,304GB)— Custom licence – read it2026-08-08
Qwen3-Coder-Next80B / 3B 256K 177 GB54.1 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial OK2026-01-30
Qwen3-Next-80B-A3B-Instruct81.3B / 3B 256K 180 GB54.9 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial OK2025-09-09
Qwen3-Next-80B-A3B-Thinking81.3B / 3B 256K 180 GB54.9 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial OK2025-09-09
Qwen3-4B-Instruct-25074.02B 256K 10.6 GB4.4 GB 8GB GPUMac 16GB Commercial OK2025-08-05
Qwen3-4B-Thinking-25074.02B 256K 10.6 GB4.4 GB 8GB GPUMac 16GB Commercial OK2025-08-05
Qwen3-Coder-30B-A3B-Instruct30.5B / 3B 256K 68.5 GB21.7 GB RTX 5090 32GBMac 64GB Commercial OK2025-07-31
Qwen3-30B-A3B-Thinking-250730.5B / 3B 256K 68.5 GB21.7 GB RTX 5090 32GBMac 64GB Commercial OK2025-07-29
Qwen3-30B-A3B-Instruct-250730.5B / 3B 256K 68.5 GB21.7 GB RTX 5090 32GBMac 64GB Commercial OK2025-07-28
Qwen3-235B-A22B-Thinking-2507235.1B / 22B 256K 519 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial OK2025-07-25
Qwen3-Coder-480B-A35B-Instruct480.2B / 35B 256K 1.06 TB323 GB 4 × RTX PRO 6000 (384GB)Mac 512GB Commercial OK2025-07-22
Qwen3-235B-A22B-Instruct-2507235.1B / 22B 256K 519 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial OK2025-07-21
Qwen3-0.6B752M 40K 3.1 GB1.9 GB 8GB GPUMac 16GB Commercial OK2025-04-27
Qwen3-1.7B2.03B 40K 5.9 GB2.8 GB 8GB GPUMac 16GB Commercial OK2025-04-27
Qwen3-4B4.02B 40K 10.6 GB4.4 GB 8GB GPUMac 16GB Commercial OK2025-04-27
Qwen3-8B8.19B 40K 19.7 GB7.2 GB 8GB GPUMac 16GB Commercial OK2025-04-27
Qwen3-14B14.8B 40K 34.3 GB11.7 GB 16GB GPUMac 24GB Commercial OK2025-04-27
Qwen3-30B-A3B30.5B / 3B 40K 68.5 GB21.7 GB RTX 5090 32GBMac 64GB Commercial OK2025-04-27
Qwen3-32B32.8B 40K 74.7 GB24.5 GB RTX 5090 32GBMac 64GB Commercial OK2025-04-27
Qwen3-235B-A22B235.1B / 22B 40K 519 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial OK2025-04-27
QwQ-32B32.8B 40K 74.7 GB24.5 GB RTX 5090 32GBMac 64GB Commercial OK2025-03-05
Qwen2.5-7B-Instruct-1M7.62B 1M 17.7 GB6.0 GB 8GB GPUMac 16GB Commercial OK2025-01-23
Qwen2.5-14B-Instruct-1M14.8B 1M 34.6 GB12.0 GB 16GB GPUMac 24GB Commercial OK2025-01-23
QwQ-32B-Preview32.8B 32K 74.7 GB24.5 GB RTX 5090 32GBMac 64GB Commercial OK2024-11-27
Qwen2.5-Coder-0.5B-Instruct494M 32K 1.7 GB0.9 GB 8GB GPUMac 16GB Commercial OK2024-11-06
Qwen2.5-Coder-3B-Instruct3.09B 32K 7.6 GB2.9 GB 8GB GPUMac 16GB Non-commercial2024-11-06
Qwen2.5-Coder-14B-Instruct14.8B 32K 34.6 GB12.0 GB 16GB GPUMac 24GB Commercial OK2024-11-06
Qwen2.5-Coder-32B-Instruct32.8B 32K 74.7 GB24.5 GB RTX 5090 32GBMac 64GB Commercial OK2024-11-06
Qwen2.5-Math-7B-Instruct7.62B 4K 17.7 GB6.0 GB 8GB GPUMac 16GB Commercial OK2024-09-19
Qwen2.5-Coder-1.5B-Instruct1.54B 32K 4.1 GB1.8 GB 8GB GPUMac 16GB Commercial OK2024-09-18
Qwen2.5-1.5B-Instruct1.54B 32K 4.1 GB1.8 GB 8GB GPUMac 16GB Commercial OK2024-09-17
Qwen2.5-3B-Instruct3.09B 32K 7.6 GB2.9 GB 8GB GPUMac 16GB Non-commercial2024-09-17
Qwen2.5-32B-Instruct32.8B 32K 74.7 GB24.5 GB RTX 5090 32GBMac 64GB Commercial OK2024-09-17
Qwen2.5-Coder-7B-Instruct7.62B 32K 17.7 GB6.0 GB 8GB GPUMac 16GB Commercial OK2024-09-17
Qwen2.5-0.5B-Instruct494M 32K 1.7 GB0.9 GB 8GB GPUMac 16GB Commercial OK2024-09-16
Qwen2.5-7B-Instruct7.62B 32K 17.7 GB6.0 GB 8GB GPUMac 16GB Commercial OK2024-09-16
Qwen2.5-14B-Instruct14.8B 32K 34.6 GB12.0 GB 16GB GPUMac 24GB Commercial OK2024-09-16
Qwen2.5-72B-Instruct72.7B 32K 163 GB51.7 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial with conditions2024-09-16
Qwen2.5-Math-1.5B1.54B 4K 4.1 GB1.8 GB 8GB GPUMac 16GB Commercial OK2024-09-16
Qwen2.5-Math-72B-Instruct ↗72.7B 4K 163 GB51.7 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial with conditions2024-09-16
Qwen2-Math-1.5B ↗1.54B 4K 4.1 GB1.8 GB 8GB GPUMac 16GB Commercial OK2024-08-08
Qwen2-7B7.62B 128K 17.7 GB6.0 GB 8GB GPUMac 16GB Commercial OK2024-06-04
Qwen2-57B-A14B-Instruct57.4B / 14B 32K 127 GB39.3 GB 2 × RTX 5090 (64GB)Mac 64GB Commercial OK2024-06-04
Qwen2-0.5B-Instruct494M 32K 1.7 GB0.9 GB 8GB GPUMac 16GB Commercial OK2024-06-03
Qwen2-1.5B-Instruct1.54B 32K 4.1 GB1.8 GB 8GB GPUMac 16GB Commercial OK2024-06-03
Qwen2-72B72.7B 128K 163 GB51.7 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial with conditions2024-05-22

02Vision-language (32)

ModelParamsContextBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3.8-Flash-Next180B / 6B 256K 397 GB121 GB H200 141GBMac 256GB Custom licence – read it2026-08-24
Qwen3.8-27B27.8B 256K 62.2 GB19.6 GB RTX 4090 24GBMac 36GB Commercial OK2026-08-05
Qwen3.6-27B27.8B 256K 62.2 GB19.6 GB RTX 4090 24GBMac 36GB Commercial OK2026-04-21
Qwen3.6-35B-A3B36B / 3B 256K 79.8 GB24.6 GB RTX 5090 32GBMac 64GB Commercial OK2026-04-15
Qwen3.5-0.8B873M 256K 2.5 GB1.2 GB 8GB GPUMac 16GB Commercial OK2026-02-28
Qwen3.5-2B2.27B 256K 5.6 GB2.1 GB 8GB GPUMac 16GB Commercial OK2026-02-28
Qwen3.5-4B4.66B 256K 11.0 GB3.9 GB 8GB GPUMac 16GB Commercial OK2026-02-27
Qwen3.5-9B9.65B 256K 22.0 GB7.2 GB 8GB GPUMac 16GB Commercial OK2026-02-27
Qwen3.5-27B27.8B 256K 62.2 GB19.6 GB RTX 4090 24GBMac 36GB Commercial OK2026-02-24
Qwen3.5-35B-A3B36B / 3B 256K 79.8 GB24.6 GB RTX 5090 32GBMac 64GB Commercial OK2026-02-24
Qwen3.5-122B-A10B125.1B / 10B 256K 276 GB84.1 GB RTX PRO 6000 96GBMac 128GB Commercial OK2026-02-24
Qwen3.5-397B-A17B403.4B / 17B 256K 888 GB270 GB 4 × RTX PRO 6000 (384GB)Mac 512GB Commercial OK2026-02-16
Qwen3-VL-2B-Instruct2.13B 256K 6.1 GB2.9 GB 8GB GPUMac 16GB Commercial OK2025-10-19
Qwen3-VL-32B-Instruct33.4B 256K 76.0 GB24.9 GB RTX 5090 32GBMac 64GB Commercial OK2025-10-19
Qwen3-VL-2B-Thinking2.13B 256K 6.1 GB2.9 GB 8GB GPUMac 16GB Commercial OK2025-10-19
Qwen3-VL-32B-Thinking33.4B 256K 76.0 GB24.9 GB RTX 5090 32GBMac 64GB Commercial OK2025-10-19
Qwen3-VL-4B-Instruct4.44B 256K 11.5 GB4.7 GB 8GB GPUMac 16GB Commercial OK2025-10-11
Qwen3-VL-8B-Instruct8.77B 256K 21.0 GB7.6 GB 12GB GPUMac 16GB Commercial OK2025-10-11
Qwen3-VL-4B-Thinking4.44B 256K 11.5 GB4.7 GB 8GB GPUMac 16GB Commercial OK2025-10-11
Qwen3-VL-8B-Thinking8.77B 256K 21.0 GB7.6 GB 12GB GPUMac 16GB Commercial OK2025-10-11
Qwen3-VL-30B-A3B-Instruct31.1B / 3B 256K 69.7 GB22.0 GB RTX 5090 32GBMac 64GB Commercial OK2025-09-30
Qwen3-VL-30B-A3B-Thinking31.1B / 3B 256K 69.7 GB22.0 GB RTX 5090 32GBMac 64GB Commercial OK2025-09-30
Qwen3-VL-235B-A22B-Instruct235.7B / 22B 256K 521 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial OK2025-09-22
Qwen3-VL-235B-A22B-Thinking235.7B / 22B 256K 521 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial OK2025-09-22
Qwen2.5-VL-32B-Instruct33.5B 125K 76.2 GB25.0 GB RTX 5090 32GBMac 64GB Commercial OK2025-03-21
Qwen2.5-VL-72B-Instruct73.4B 125K 165 GB52.1 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial with conditions2025-01-27
Qwen2.5-VL-3B-Instruct3.75B 125K 9.1 GB3.3 GB 8GB GPUMac 16GB Non-commercial2025-01-26
Qwen2.5-VL-7B-Instruct8.29B 125K 19.2 GB6.5 GB 8GB GPUMac 16GB Commercial OK2025-01-26
QVQ-72B-Preview ↗73.4B 125K 165 GB52.1 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial with conditions2024-12-24
Qwen2-VL-72B-Instruct73.4B 32K 165 GB52.1 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial with conditions2024-09-17
Qwen2-VL-2B-Instruct2.21B 32K 5.6 GB2.2 GB 8GB GPUMac 16GB Commercial OK2024-08-28
Qwen2-VL-7B-Instruct8.29B 32K 19.2 GB6.5 GB 8GB GPUMac 16GB Commercial OK2024-08-28

03Omni / audio LLM (5)

ModelParamsContextBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3-Omni-30B-A3B-Instruct35.3B / 3B 64K 78.9 GB24.8 GB RTX 5090 32GBMac 64GB Commercial OK2025-09-20
Qwen3-Omni-30B-A3B-Thinking31.7B / 3B 64K 71.1 GB22.5 GB RTX 5090 32GBMac 64GB Commercial OK2025-09-15
Qwen2.5-Omni-3B5.54B 32K 13.0 GB4.5 GB 8GB GPUMac 16GB Non-commercial2025-04-30
Qwen2.5-Omni-7B10.7B 32K 24.6 GB8.1 GB 12GB GPUMac 16GB Commercial OK2025-03-22
Qwen2-Audio-7B-Instruct8.4B — 19.0 GB6.1 GB 8GB GPUMac 16GB Commercial OK2024-07-31

04Embedding (5)

ModelParamsBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3-VL-Embedding-2B2.13B 5.2 GB1.9 GB 8GB GPUMac 16GB Commercial OK2026-01-07
Qwen3-VL-Embedding-8B8.14B 18.4 GB5.9 GB 8GB GPUMac 16GB Commercial OK2026-01-07
Qwen3-Embedding-0.6B596M 1.8 GB0.9 GB 8GB GPUMac 16GB Commercial OK2025-06-03
Qwen3-Embedding-4B4.02B 9.3 GB3.2 GB 8GB GPUMac 16GB Commercial OK2025-06-03
Qwen3-Embedding-8B7.57B 17.1 GB5.5 GB 8GB GPUMac 16GB Commercial OK2025-06-03

05Reranker (5)

ModelParamsBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3-VL-Reranker-2B2.13B 5.2 GB1.9 GB 8GB GPUMac 16GB Commercial OK2026-01-07
Qwen3-VL-Reranker-8B8.77B 19.8 GB6.3 GB 8GB GPUMac 16GB Commercial OK2026-01-07
Qwen3-Reranker-4B4.02B 9.3 GB3.2 GB 8GB GPUMac 16GB Commercial OK2025-06-03
Qwen3-Reranker-0.6B596M 1.8 GB0.9 GB 8GB GPUMac 16GB Commercial OK2025-05-29
Qwen3-Reranker-8B8.19B 18.5 GB6.0 GB 8GB GPUMac 16GB Commercial OK2025-05-29

06Speech-to-text (2)

ModelParamsBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3-ASR-0.6B938M 2.6 GB1.1 GB 8GB GPUMac 16GB Commercial OK2026-01-28
Qwen3-ASR-1.7B2.35B 5.7 GB2.1 GB 8GB GPUMac 16GB Commercial OK2026-01-28

07Text-to-speech / audio (5)

ModelParamsBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3-TTS-12Hz-0.6B-Base915M 2.5 GB1.1 GB 8GB GPUMac 16GB Commercial OK2026-01-21
Qwen3-TTS-12Hz-1.7B-Base1.93B 4.7 GB1.8 GB 8GB GPUMac 16GB Commercial OK2026-01-21
Qwen3-TTS-12Hz-0.6B-CustomVoice906M 2.5 GB1.1 GB 8GB GPUMac 16GB Commercial OK2026-01-21
Qwen3-TTS-12Hz-1.7B-CustomVoice1.92B 4.7 GB1.8 GB 8GB GPUMac 16GB Commercial OK2026-01-21
Qwen3-TTS-12Hz-1.7B-VoiceDesign1.92B 4.7 GB1.8 GB 8GB GPUMac 16GB Commercial OK2026-01-21

08Image generation (8)

ModelParamsBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen-Image-2.1-Turbo7.12B 16.2 GB5.2 GB 8GB GPUMac 16GB Non-commercial2026-10-09
Qwen-Image-2.17.12B 16.2 GB5.2 GB 8GB GPUMac 16GB Non-commercial2026-09-14
Qwen-Image-251220.4B 45.4 GB14.1 GB 16GB GPUMac 24GB Commercial OK2025-12-30
Qwen-Image-Edit-251120.4B 45.4 GB14.1 GB 16GB GPUMac 24GB Commercial OK2025-12-17
Qwen-Image-Layered20.4B 45.4 GB14.1 GB 16GB GPUMac 24GB Commercial OK2025-12-17
Qwen-Image-Edit-250920.4B 45.4 GB14.1 GB 16GB GPUMac 24GB Commercial OK2025-09-22
Qwen-Image-Edit20.4B 45.4 GB14.1 GB 16GB GPUMac 24GB Commercial OK2025-08-17
Qwen-Image20.4B 45.4 GB14.1 GB 16GB GPUMac 24GB Commercial OK2025-08-02

09Safety / guard (4)

ModelParamsContextBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
Qwen3Guard-Gen-0.6B752M 32K 3.1 GB1.9 GB 8GB GPUMac 16GB Commercial OK2025-09-23
Qwen3Guard-Gen-4B4.41B 32K 11.4 GB4.7 GB 8GB GPUMac 16GB Commercial OK2025-09-23
Qwen3Guard-Gen-8B8.19B 32K 19.7 GB7.2 GB 8GB GPUMac 16GB Commercial OK2025-09-23
Qwen3Guard-Stream-0.6B597M 8K 2.8 GB1.8 GB 8GB GPUMac 16GB Commercial OK2025-09-23

10Other families

IBM Granite · NVIDIA Nemotron · Gemma · Mistral / Magistral / Devstral · GLM / CogVideoX (Z.ai) · DeepSeek · Ai2 OLMo / Molmo · Ant Ling / Ring / Ming · Liquid LFM · MiniCPM / VoxCPM · Tencent Hunyuan / Hy · Phi · Llama · Meituan LongCat · Jina embeddings & rerankers · Wan · Kimi · Cohere Command / Aya · Xiaomi MiMo · FLUX · Microsoft VibeVoice / Fara / Harrier · InternLM / Intern-S · StepFun Step · MiniMax · Baidu ERNIE · NVIDIA Parakeet / Canary speech · Stable Diffusion / Stable Audio · Yi (01.AI) · ByteDance Seed · SmolLM / SmolVLM · BGE embeddings & rerankers · LTX Video · FunAudioLLM speech · gpt-oss · poolside Laguna · Whisper · Baichuan · Thinking Machines Inkling · xAI Grok · rednote dots

Totals include 10% + 0.5GB overhead and one 8K-token KV cache for text models. Estimates only. Method →

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