Mistral Small 4 VRAM requirements
Mistral · 119B params (6.5B active) · Apache 2.0. Estimates, not vendor numbers.
Mistral Small 4 (119B parameters, 6.5B active per token) needs about 73 GB of VRAM or unified memory at Q4, 86 GB at Q5 and 128 GB at Q8 (weights plus about 1.5 GB headroom, our estimate). The smallest common setup that runs it is Mac 128GB (M4 Max).
| Quant | Download | Memory needed |
|---|---|---|
| Q4 | 71 GB | 73 GB |
| Q5 | 85 GB | 86 GB |
| Q8 | 126 GB | 128 GB |
Will it run on my machine?
| Setup | Runs? | Speed est. |
|---|---|---|
| 8GB GPU (RTX 4060) | No | n/a |
| 12GB GPU (RTX 3060 / 4070) | No | n/a |
| 16GB GPU (RTX 4060 Ti 16GB) | No | n/a |
| 24GB GPU (RTX 4090) | No | n/a |
| 32GB GPU (RTX 5090) | No | n/a |
| Mac 16GB (M4) | No | n/a |
| Mac 48GB (M4 Pro) | No | n/a |
| Mac 128GB (M4 Max) | Yes, Q5 | ~71 tok/s |
| PC, 32GB RAM, no GPU | No | n/a |
Best quant that fits each setup, with estimated generation speed from memory bandwidth. Real speed depends on the runtime, context length and settings.
Mixture of experts: 119B total, 6.5B active (Mistral).
FAQ
How much VRAM does Mistral Small 4 need?
About 73 GB at Q4, 86 GB at Q5 and 128 GB at Q8, including about 1.5 GB headroom. These are estimates from the parameter count on the official model card.
Can Mistral Small 4 run on 8GB of VRAM?
Not fully on the GPU. At Q4 it needs about 73 GB; partial CPU offload works but is much slower.
Can Mistral Small 4 run on a Mac?
Yes. Mac 128GB (M4 Max) can run it.
Which quant of Mistral Small 4 should I use?
Use the highest quant that fits with room to spare: Q8 is near full quality, Q5 is a good balance, and Q4 is the smallest that usually keeps quality acceptable.
Why is Mistral Small 4 fast for its size?
It is a mixture of experts: 119B parameters in total but only 6.5B are used per token, so it reads less memory per token. You still need memory for all 119B.
Similar-size models
Gemma 4 31B (30.7B) · DeepSeek R1 Distill Qwen 32B (32.8B) · Qwen3.6 35B A3B (35B) · Llama 3.3 70B Instruct (70B) · gpt-oss-120b (117B) · Qwen3.8 Flash Next (180B)