A smaller variant of O3, expected to offer enhanced multimodal capabilities, improved reasoning, and more efficient resource utilization compared to previous models while maintaining strong performance on core tasks.
发布日期2025年1月30日
参数规模—
上下文长度200K
许可证Proprietary
知识截止2023年9月30日
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| COLLIE | language reasoning writing | 98.7 | 来源 |
| MATH | math reasoning | 97.9 | 来源 |
| IFEval | general | 93.9 | 来源 |
| MGSM | math reasoning | 92.0 | 来源 |
| AIME 2024 | math reasoning | 87.3 | 来源 |
| MMLU | general reasoning language math | 86.9 | 来源 |
| LiveBench | math reasoning general | 84.6 | 来源 |
| Multilingual MMLU | general reasoning language | 80.7 | 来源 |
| Multi-IF | reasoning communication language | 79.5 | 来源 |
| GPQA | reasoning general | 77.2 | 来源 |
| Aider-Polyglot | general code | 66.7 | 来源 |
| Aider-Polyglot Edit | general code | 60.4 | 来源 |
| Graphwalks parents <128k | reasoning spatial_reasoning | 58.3 | 来源 |
| TAU-bench Retail | reasoning communication | 57.6 | 来源 |
| Graphwalks BFS <128k | reasoning spatial_reasoning | 51.0 | 来源 |
| MultiChallenge (o3-mini grader) | reasoning language | 50.2 | 来源 |
| Internal API instruction following (hard) | general | 50.0 | 来源 |
| SWE-Bench Verified | reasoning frontend_development code | 49.3 | 来源 |
| MultiChallenge | communication reasoning | 39.9 | 来源 |
| TAU-bench Airline | reasoning communication | 32.4 | 来源 |
| OpenAI-MRCR: 2 needle 128k | long_context reasoning | 18.7 | 来源 |
| SWE-Lancer | reasoning code | 18.0 | 来源 |
| ComplexFuncBench | long_context reasoning | 17.6 | 来源 |
| SimpleQA | general reasoning | 15.0 | 来源 |
| FrontierMath | math reasoning | 9.2 | 来源 |
| SWE-Lancer (IC-Diamond subset) | reasoning code | 7.4 | 来源 |
Pricing
API 价格对比
| 服务商 | 输入价 | 输出价 | 上下文 | 吞吐(tok/s) | 延迟(s) | 函数调用 | 代码执行 | 联网搜索 |
|---|---|---|---|---|---|---|---|---|
| Abacus | $1.10 | $4.40 | 200K | — | — | ✓ | ✗ | ✗ |
| Azure | $1.10 | $4.40 | 200K | — | — | ✓ | ✗ | ✗ |
| Azure Cognitive Services | $1.10 | $4.40 | 200K | — | — | ✓ | ✗ | ✗ |
| Helicone | $1.10 | $4.40 | 200K | — | — | ✓ | ✗ | ✗ |
| Jiekou.AI | $1.10 | $4.40 | 131K | — | — | ✓ | ✗ | ✗ |
| LLM Gateway | $1.10 | $4.40 | 200K | — | — | ✓ | ✗ | ✗ |
| OpenAI | $1.10 | $4.40 | 200K | — | — | ✓ | ✗ | ✗ |
价格单位：美元/百万 token，数据来自社区整理，仅供参考。