Building upon Mistral Small 3 (2501), Mistral Small 3.1 (2503) adds state-of-the-art vision understanding and enhances long context capabilities up to 128k tokens without compromising text performance. With 24 billion parameters, this model achieves top-tier capabilities in both text and vision tasks.
发布日期2025年3月17日
参数规模24B
上下文长度—
许可证Apache 2.0
知识截止—
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| HumanEval | reasoning code | 88.4 | 来源 |
| MMLU | general reasoning language math | 80.6 | 来源 |
| TriviaQA | general reasoning | 80.5 | 来源 |
| MBPP | reasoning general | 74.7 | 来源 |
| MATH | math reasoning | 69.3 | 来源 |
| MMLU-Pro | language reasoning math general | 66.8 | 来源 |
| MMMU | multimodal reasoning general | 59.3 | 来源 |
| GPQA | reasoning general | 46.0 | 来源 |
| SimpleQA | general reasoning | 10.4 | 来源 |
Pricing
API 价格对比
暂无 API 价格。