Mistral Large 2
Mistral·
LLMsopen-weightWorkstation
Overview
123B parameter open model with strong coding capabilities. Competitive with GPT-4o and Claude 3.5 Sonnet on code generation tasks.
Capabilities and innovations
123B parameters128K context window80+ coding languagesFunction calling and JSON modeInstruction-following improvementsReduced hallucination rateImproved multi-turn conversation
Benchmarks
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 46.6% | unsourced |
| MMLU | 84% | unsourced |
Architecture and hardware
- Parameters
- 123B total (Dense)
- Estimated VRAM at Q4
- ~71 GB, Workstation class
- Quantization formats
- GGUF, GPTQ, AWQ
- Recommended runtime
- vLLM
- License
- Mistral Research License
Links
More from Mistral
Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.