// Open-model VRAM table · family
NVIDIA Nemotron: local deployment & VRAM
62 open-weight models from NVIDIA (108 official repos incl. quantised builds). At 4-bit they range from 0.9 GB (nemocurator-fineweb-nemotron-4-edu-classifier) to 374 GB (NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16).
official orgs: nvidialatest 2026-10-16memory = estimate
01Text LLM (37)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 | 3.98B | 256K | 9.7 GB | 3.6 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-10-16 |
| Nemotron-3-Labs-Ultra-Math-RL | 550B / 55B | 256K | 1.21 TB | 367 GB | 8 × H100 (640GB) | — | Commercial OK | 2026-09-02 |
| NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 | 31.6B / 3B | 256K | 70.0 GB | 21.6 GB | RTX 4090 24GB | Mac 36GB | Commercial OK | 2026-08-01 |
| NVIDIA-NemotronLabs-VoiceChat-11B | 11.1B | — | 24.9 GB | 7.9 GB | 12GB GPU | Mac 16GB | Commercial OK | 2026-07-29 |
| Nemotron-Labs-Audex-2B ↗ | 2B | — | 4.9 GB | 1.8 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-07-06 |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | 560.5B / 55B | 256K | 1.23 TB | 374 GB | 8 × H100 (640GB) | — | Commercial OK | 2026-06-03 |
| Nemotron-Labs-Diffusion-14B | 13.5B | 256K | 31.6 GB | 10.8 GB | 12GB GPU | Mac 24GB | Commercial with conditions | 2026-04-22 |
| Nemotron-Cascade-2-30B-A3B | 31.6B / 3B | 256K | 70.0 GB | 21.6 GB | RTX 4090 24GB | Mac 36GB | Commercial with conditions | 2026-03-18 |
| Nemotron-Labs-Diffusion-8B | 8.49B | 256K | 20.3 GB | 7.3 GB | 12GB GPU | Mac 16GB | Commercial with conditions | 2026-03-18 |
| NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 123.6B / 12B | 256K | 273 GB | 83.0 GB | RTX PRO 6000 96GB | Mac 128GB | Commercial with conditions | 2026-03-10 |
| NVIDIA-Nemotron-3-Nano-4B-BF16 | 3.97B | 256K | 9.4 GB | 3.3 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2026-03-07 |
| Nemotron-Labs-Diffusion-3B | 3.83B | 256K | 9.8 GB | 3.9 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2026-03-02 |
| NVIDIA-Nemotron-Nano-9B-v2-Japanese | 8.89B | 128K | 20.2 GB | 6.6 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2026-02-04 |
| Nemotron-Cascade-8B-Thinking ↗ | 8.19B | 32K | 19.7 GB | 7.2 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-12-08 |
| Nemotron-Cascade-14B-Thinking ↗ | 14.8B | 32K | 34.3 GB | 11.7 GB | 16GB GPU | Mac 24GB | Commercial with conditions | 2025-12-08 |
| NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 31.6B / 3B | 256K | 70.0 GB | 21.6 GB | RTX 4090 24GB | Mac 36GB | Commercial with conditions | 2025-12-04 |
| Nemotron-Orchestrator-8B ↗ | 8.19B | 40K | 19.7 GB | 7.2 GB | 8GB GPU | Mac 16GB | Licence not stated | 2025-11-25 |
| Nemotron-Flash-1B ↗ | 965M | 35K | 3.0 GB | 1.5 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-11-06 |
| NVIDIA-Nemotron-Nano-12B-v2 | 12.3B | 128K | 27.8 GB | 8.9 GB | 12GB GPU | Mac 16GB | Commercial with conditions | 2025-08-21 |
| NVIDIA-Nemotron-Nano-9B-v2 | 8.89B | 128K | 20.2 GB | 6.6 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-08-12 |
| Llama-3_3-Nemotron-Super-49B-v1_5 | 49.9B | 128K | 112 GB | 35.4 GB | 2 × RTX 5090 (64GB) | Mac 64GB | Commercial with conditions | 2025-07-25 |
| OpenReasoning-Nemotron-7B ↗ | 7.62B | 128K | 17.7 GB | 6.0 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-07-15 |
| Nemotron-H-8B-Reasoning-128K ↗ | 8.1B | 128K | 18.5 GB | 6.0 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-06-05 |
| Llama-3.1-Nemotron-Nano-4B-v1.1 | 4.51B | 128K | 11.5 GB | 4.6 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-05-03 |
| Nemotron-H-4B-Instruct-128K ↗ | 4.49B | 128K | 10.5 GB | 3.6 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-04-15 |
| Nemotron-H-56B-Base-8K | 56.3B | 8K | 125 GB | 38.4 GB | 2 × RTX 5090 (64GB) | Mac 64GB | Non-commercial | 2025-04-08 |
| Llama-3_1-Nemotron-Ultra-253B-v1 | 253.4B | 128K | 560 GB | 172 GB | 4 × RTX PRO 6000 (384GB) | Mac 256GB | Commercial with conditions | 2025-04-07 |
| Nemotron-H-4B-Base-8K | 4.49B | 8K | 10.5 GB | 3.6 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-03-20 |
| Nemotron-H-8B-Base-8K | 8.1B | 8K | 18.5 GB | 6.0 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-03-19 |
| Llama-3.1-Nemotron-Nano-8B-v1 | 8.03B | 128K | 19.2 GB | 6.9 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-03-16 |
| Llama-3_3-Nemotron-Super-49B-v1 | 49.9B | 128K | 112 GB | 35.4 GB | 2 × RTX 5090 (64GB) | Mac 64GB | Commercial with conditions | 2025-03-16 |
| Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct ↗ | 8.04B | 1M | 19.3 GB | 6.9 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-03-04 |
| nemocurator-fineweb-nemotron-4-edu-classifier ↗ | 109M | 1K | 1.0 GB | 0.9 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2025-02-07 |
| Nemotron-4-Mini-Hindi-4B-Instruct ↗ | 4.19B | 4K | 10.8 GB | 4.4 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2024-10-22 |
| Llama-3.1-Nemotron-70B-Instruct-HF | 70.6B | 128K | 158 GB | 50.2 GB | 2 × RTX 5090 (64GB) | Mac 128GB | Commercial with conditions | 2024-10-12 |
| Llama-3_1-Nemotron-51B-Instruct ↗ | 51.5B | 128K | 115 GB | 36.4 GB | 2 × RTX 5090 (64GB) | Mac 64GB | Commercial with conditions | 2024-09-22 |
| Nemotron-Mini-4B-Instruct | 4B | 4K | 10.4 GB | 4.2 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2024-09-10 |
02Vision-language (10)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 | 124.5B / 12B | — | 274 GB | 83.5 GB | RTX PRO 6000 96GB | Mac 128GB | Commercial OK | 2026-10-12 |
| NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16 | 74.9B / 9B | — | 165 GB | 50.4 GB | 2 × RTX 5090 (64GB) | Mac 128GB | Commercial OK | 2026-10-09 |
| Nemotron-3.5-Parse | 631M | 16K | 2.1 GB | 1.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-09-03 |
| NVIDIA-Nemotron-Parse-2.0 | 903M | — | 2.5 GB | 1.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-06-30 |
| Nemotron-Labs-Diffusion-VLM-8B ↗ | 8.92B | 256K | 21.3 GB | 7.6 GB | 12GB GPU | Mac 16GB | Non-commercial | 2026-05-08 |
| nemotron-ocr-v2 ↗ | — | — | — | — | — | — | Commercial with conditions | 2026-04-01 |
| NVIDIA-Nemotron-Parse-v1.2 | 936M | — | 2.6 GB | 1.1 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2026-02-18 |
| NVIDIA-Nemotron-Parse-v1.1 | 957M | — | 2.6 GB | 1.1 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-11-15 |
| NVIDIA-Nemotron-Nano-12B-v2-VL-BF16 | 13.2B | 128K | 29.7 GB | 9.5 GB | 12GB GPU | Mac 16GB | Commercial with conditions | 2025-10-21 |
| Llama-3.1-Nemotron-Nano-VL-8B-V1 | 8.72B | 128K | 20.8 GB | 7.4 GB | 12GB GPU | Mac 16GB | Commercial with conditions | 2025-06-03 |
03Omni / audio LLM (1)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16 | 33B / 3B | 256K | 73.2 GB | 22.6 GB | RTX 5090 32GB | Mac 64GB | Commercial with conditions | 2026-04-20 |
04Embedding (11)
| Model | Params | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|
| Nemotron-3.5-Embed-1B-BF16 | 1.55B | 3.9 GB | 1.5 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-10-07 |
| Nemotron-3.5-Embed-8B-BF16 | 8.38B | 18.9 GB | 6.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-10-07 |
| Nemotron-3-Embed-1B-BF16 | 1.14B | 3.0 GB | 1.3 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-07-14 |
| Nemotron-3-Embed-8B-BF16 | 7.95B | 18.0 GB | 5.8 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-07-14 |
| nemotron-colembed-vl-4b-v2 ↗ | 4.83B | 11.1 GB | 3.7 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-01-15 |
| nemotron-colembed-vl-8b-v2 | 8.77B | 19.8 GB | 6.3 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-01-15 |
| llama-nemotron-colembed-vl-3b-v2 | 4.41B | 10.2 GB | 3.4 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-01-14 |
| llama-nemotron-embed-vl-1b-v2 | 1.68B | 4.2 GB | 1.6 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-12-03 |
| llama-nemotron-embed-1b-v2 | 1.24B | 3.2 GB | 1.3 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-10-16 |
| llama-embed-nemotron-8b | 7.5B | 17.0 GB | 5.5 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-10-07 |
| omni-embed-nemotron-3b ↗ | 4.7B | 10.8 GB | 3.6 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-09-30 |
05Reranker (3)
| Model | Params | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|
| Nemotron-3.5-Rerank-8B-BF16 | 8.38B | 18.9 GB | 6.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-10-15 |
| llama-nemotron-rerank-vl-1b-v2 | 1.68B | 4.2 GB | 1.6 GB | 8GB GPU | Mac 16GB | Commercial with conditions | 2025-12-04 |
| llama-nemotron-rerank-1b-v2 | 1.24B | 3.2 GB | 1.3 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-10-16 |
06Other families
Qwen · IBM Granite · 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 →