Mistral 7B v0.3 hardware requirements
7.25B parameters. Needs about 6.1 GB at Q4_K_M with 8K context. Model card.
Memory needed at each quantization
| Quant | Quality | 8K context | 32K context |
|---|---|---|---|
| FP16 | Full quality | 16.2 GB | 19.4 GB |
| Q8_0 | Near-lossless | 9.4 GB | 12.6 GB |
| Q6_K | Excellent | 7.6 GB | 10.8 GB |
| Q5_K_M | Very good | 6.8 GB | 10.1 GB |
| Q4_K_M | Good, most popular | 6.1 GB | 9.3 GB |
| Q3_K_M | Noticeably degraded | 5.2 GB | 8.4 GB |
| Q2_K | Heavily degraded | 4.4 GB | 7.6 GB |
Run it with Ollama: ollama run mistral:7b (downloads Q4_K_M by default).
Run it in the cloud
Don't want to buy hardware? Run Mistral 7B v0.3 on a rented GPU today.
Rent a cloud GPU by the hour. No upfront cost, cancel any time.
Every GPU and Mac checked
| GPU | Runs? | Best fit |
|---|---|---|
| NVIDIA RTX 3060 12GB | Yes | Q8_0 · ~9.4 GB · ~28 tok/s |
| NVIDIA RTX 3070 | Yes | Q6_K · ~7.6 GB · ~45 tok/s |
| NVIDIA RTX 3080 10GB | Yes | Q8_0 · ~9.4 GB · ~59 tok/s |
| NVIDIA RTX 3090 | Yes | FP16 · ~16.2 GB · ~39 tok/s |
| NVIDIA RTX 4060 | Yes | Q6_K · ~7.6 GB · ~27 tok/s |
| NVIDIA RTX 4060 Ti 16GB | Yes | Q8_0 · ~9.4 GB · ~22 tok/s |
| NVIDIA RTX 5060 Ti 16GB | Yes | Q8_0 · ~9.4 GB · ~35 tok/s |
| NVIDIA RTX 4070 | Yes | Q8_0 · ~9.4 GB · ~39 tok/s |
| NVIDIA RTX 4070 Ti Super | Yes | Q8_0 · ~9.4 GB · ~52 tok/s |
| NVIDIA RTX 4080 | Yes | Q8_0 · ~9.4 GB · ~56 tok/s |
| NVIDIA RTX 4090 | Yes | FP16 · ~16.2 GB · ~42 tok/s |
| NVIDIA RTX 5070 | Yes | Q8_0 · ~9.4 GB · ~52 tok/s |
| NVIDIA RTX 5070 Ti | Yes | Q8_0 · ~9.4 GB · ~70 tok/s |
| NVIDIA RTX 5080 | Yes | Q8_0 · ~9.4 GB · ~75 tok/s |
| NVIDIA RTX 5090 | Yes | FP16 · ~16.2 GB · ~74 tok/s |
| AMD Radeon RX 7800 XT | Yes | Q8_0 · ~9.4 GB · ~49 tok/s |
| AMD Radeon RX 7900 XTX | Yes | FP16 · ~16.2 GB · ~40 tok/s |
| Apple M4 (16GB) | Yes | Q8_0 · ~9.4 GB · ~9 tok/s |
| Apple M1 (16GB) | Yes | Q8_0 · ~9.4 GB · ~5 tok/s |
| Apple M2 Pro (32GB) | Yes | FP16 · ~16.2 GB · ~8 tok/s |
| Apple M3 Max (64GB) | Yes | FP16 · ~16.2 GB · ~17 tok/s |
| Apple M4 Pro (48GB) | Yes | FP16 · ~16.2 GB · ~11 tok/s |
| Apple M4 Max (128GB) | Yes | FP16 · ~16.2 GB · ~23 tok/s |
| Apple M2 Ultra (192GB) | Yes | FP16 · ~16.2 GB · ~33 tok/s |