Mistral Large 2
Mistral·
LLMsflagshipcloud + localWorkstation
Overview
Significant improvements in code generation and reasoning.
Capabilities and innovations
123B ParametersCode generationFunction calling123B ParametersCode-focused TrainingAdvanced Function Calling
Benchmarks
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 55.1% | unsourced |
| Humanity’s Last Exam (no tools) | 8.9% | unsourced |
| SWE-bench Verified | 32.6% | unsourced |
| MMLU | 84% | unsourced |
- SWE-bench Pro: benchmark did not exist at release
- MMLU-Pro: benchmark did not exist at release
- Terminal-Bench 2.x: benchmark did not exist at release
- MMMU-Pro: benchmark did not exist at release
Architecture and hardware
- Parameters
- 123B total (Dense)
- Estimated VRAM at Q4
- ~71 GB, Workstation class
Links
More from Mistral
Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.