Mistral 7B Instruct V0.2
mistralai · released 2023-12-11 · apache-2.0 license
Mistral-7B-Instruct-v0.2 (Dec 2023, Apache-2.0) is an older 7B chat model (32K context). Open LLM Leaderboard records MMLU-Pro 19.08 and IFEval 54.96; widely adopted but now superseded.
Key specs
| Type | Local open-weight |
|---|---|
| Parameters | 7.24B total |
| Architecture | mistral |
| Context window | 33K tokens |
| Knowledge cutoff | — |
| Modalities | text |
| Recommended backends | — |
| Minimum viable rig | — |
Benchmark scores
| GPQA Diamond | — |
|---|---|
| SWE-bench Verified | — |
| AIME | — |
| MMLU-Pro | 19.08% |
| BFCL v3 (tool use) | — |
| Composite score | 3.39 |
| Community rating | No reviews yet |
VRAM & disk per quantization
| Quant | VRAM | Disk | RAM | Context |
|---|---|---|---|---|
| Q4_K_M | 5.7 GB | 4.2 GB | — | 33K |
Strengths & weaknesses
Strengths: Strong instruction-following / chat for a 2023 7B; 32K context; Apache-2.0, quantization-friendly
Weaknesses: Weak on hard knowledge/reasoning (MMLU-Pro 19.08); Largely superseded by Mistral-Nemo and newer 7Bs