Mistral: Mistral Large 3 2512
● Excellent8.5
/ 10
Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.
Specifications
| Attribute | Value |
|---|---|
| Lab | Mistral AI |
| Tags | Intelligent Coding |
| Overall Score | 8.5/10 |
| Release Date | 2025-12 |
| Context Window | 262,144 tokens |
| Input Price / 1M | $0.50 |
| Output Price / 1M | $1.50 |
| Input Modalities | Text, Image, File |
| Output Modalities | Text |
Strengths
- Updated flagship with improved reasoning
- Strong multilingual performance
- Agentic workflow capabilities
- Good coding and tool use
Weaknesses
- Behind top-tier flagships on hard reasoning
- Moderate context window vs peers
Best For
- Enterprise reasoning tasks
- Multilingual applications
- Agentic coding workflows
Sources & Further Reading
Related Models
Scores are aggregated from public benchmarks (MMLU, HumanEval, MATH, GSM8K, LMSYS) and normalized to a 1–10 scale. Methodology →