OpenAI o3-mini (medium)
OpenAI o3-mini (medium) is an AI model published by OpenAI, released on 2025-01-31, for 推理大模型, and 128k tokens context length, under the 不开源 license.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
OpenAI o3-mini (medium) currently shows benchmark results led by LiveCodeBench (46 / 109, score 67.40), FrontierMath (26 / 54, score 8.10). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.
OpenAI o3 mini是OpenAI发布的第二代推理大模型中最小规模的版本。第一代推理大模型是o1，跳过了o2版本号的原因可能是o2与英国电信运营商的品牌冲突。
o3 mini总共分为三个版本，o3-mini low，o3-mini medium 和o3-mini high。本模型为其中中等规模的版本，即o3-mini medium。ChatGPT中免费用户使用的是o3-mini meidum版本。
不同版本的评测结果如下：
| Category | 01-mini | 03-mini (low) | 03-mini (medium) | 03-mini (high) |
|---|---|---|---|---|
| General | 85.2 | 84.9 | 85.9 | 86.9 |
| Math | 90.0 | 95.8 | 97.3 | 97.9 |
| Factuality | 89.9 | 55.1 | 90.8 | 92.0 |
| Eval | 76.0 | 13.0 | 13.4 | 13.8 |
| Model | Reasoning Level | Average | LCB Generation | Code Completion |
| o3-mini | low | 0.618 | 0.756 | 0.48 |
| o3-mini | medium | 0.723 | 0.846 | 0.60 |
| o3-mini | high | 0.846 | 0.820 | 0.833 |
| o1-mini | high | 0.674 | 0.628 | 0.72 |
欢迎关注 DataLearner 官方微信，获得最新 AI 技术推送
