A text-only model that delivers lowest-latency responses at very low cost while maintaining strong performance on core language tasks. Optimized for speed and efficiency while preserving high accuracy on key benchmarks.
发布日期2024年11月20日
参数规模—
上下文长度—
许可证Proprietary
知识截止—
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| GSM8k | math reasoning | 92.3 | 来源 |
| ARC-C | reasoning general | 90.2 | 来源 |
| Translation Set1→en COMET22 | language | 88.7 | 来源 |
| Translation en→Set1 COMET22 | language | 88.5 | 来源 |
| IFEval | general | 87.2 | 来源 |
| HumanEval | reasoning code | 81.1 | 来源 |
| BBH | reasoning math language | 79.5 | 来源 |
| DROP | reasoning math | 79.3 | 来源 |
| MMLU | general reasoning language math | 77.6 | 来源 |
| MATH | math reasoning | 69.3 | 来源 |
| FinQA | finance math reasoning | 65.2 | 来源 |
| BFCL | general reasoning | 56.2 | 来源 |
| CRAG | reasoning search | 43.1 | 来源 |
| Translation Set1→en spBleu | language | 42.6 | 来源 |
| Translation en→Set1 spBleu | language | 40.2 | 来源 |
| GPQA | reasoning general | 40.0 | 来源 |
| SQuALITY | summarization long_context language | 18.8 | 来源 |
Pricing
API 价格对比
暂无 API 价格。