// Open-model VRAM table · family
Cohere Command / Aya: local deployment & VRAM
14 open-weight models from Cohere (19 official repos incl. quantised builds). At 4-bit they range from 2.7 GB (North-Micro-Vision-Instruct) to 147 GB (command-a-plus-05-2026-bf16).
official orgs: CohereLabslatest 2026-08-10memory = estimate
01Text LLM (10)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| North-Mini-Code-1.0 | 30B / 3B | 488K | 67.0 GB | 21.0 GB | RTX 4090 24GB | Mac 36GB | Commercial OK | 2026-06-05 |
| tiny-aya-base ↗ | 3.35B | — | 7.9 GB | 2.7 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-02-13 |
| tiny-aya-global | 3.35B | — | 7.9 GB | 2.7 GB | 8GB GPU | Mac 16GB | Non-commercial | 2026-02-13 |
| command-a-reasoning-08-2025 | 111.1B | — | 245 GB | 74.6 GB | RTX PRO 6000 96GB | Mac 128GB | Non-commercial | 2025-08-12 |
| c4ai-command-a-03-2025 | 111.1B | — | 245 GB | 74.6 GB | RTX PRO 6000 96GB | Mac 128GB | Non-commercial | 2025-03-11 |
| c4ai-command-r7b-arabic-02-2025 ↗ | 8.03B | — | 18.2 GB | 5.9 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-02-27 |
| c4ai-command-r7b-12-2024 | 8.03B | — | 18.2 GB | 5.9 GB | 8GB GPU | Mac 16GB | Non-commercial | 2024-12-11 |
| aya-expanse-8b | 8.03B | — | 18.2 GB | 5.9 GB | 8GB GPU | Mac 16GB | Non-commercial | 2024-10-23 |
| aya-expanse-32b | 32.3B | — | 71.6 GB | 22.0 GB | RTX 5090 32GB | Mac 64GB | Non-commercial | 2024-10-23 |
| aya-23-8B | 8.03B | — | 18.2 GB | 5.9 GB | 8GB GPU | Mac 16GB | Non-commercial | 2024-05-19 |
02Vision-language (4)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| North-Micro-Vision-Instruct | 2.48B | 488K | 6.6 GB | 2.7 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-08-10 |
| command-a-plus-05-2026-bf16 | 218B / 25B | 195K | 481 GB | 147 GB | 4 × RTX PRO 6000 (384GB) | Mac 256GB | Commercial OK | 2026-05-11 |
| command-a-vision-07-2025 | 111.9B | — | 247 GB | 75.1 GB | RTX PRO 6000 96GB | Mac 128GB | Non-commercial | 2025-07-28 |
| aya-vision-8b ↗ | 8.63B | — | 19.5 GB | 6.3 GB | 8GB GPU | Mac 16GB | Non-commercial | 2025-03-02 |
03Other families
Qwen · 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 · 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 →