GPT-5 nano · τ²-bench(Telecom section) 35.5% · 主文证据(ev-gpt5-nano-tau2-telecom)
GPT-5 nano 的 τ²-bench · Telecom section 官方出处。GPT-5 nano (high) · 35.5%。见《Introducing GPT-5 for developers | OpenAI》Detailed benchmarks / τ²-bench telecom。证据 ev-gpt5-nano-tau2-telecom。
原文与定位
GPT-5 nano · τ²-bench(Telecom section),GPT-5 nano (high)配置,Success rate 35.5%(成绩公布 2025-08-07)。见《Introducing GPT-5 for developers | OpenAI》Detailed benchmarks / τ²-bench telecom。 查看成绩条件
Tau 2-bench telecom: GPT-5 (high) 96.7%; GPT-5 mini (high) 74.1%; GPT-5 nano (high) 35.5%.
- 官方材料
- Introducing GPT-5 for developers | OpenAI
- 定位
- 主文 · Detailed benchmarks / τ²-bench telecom
- 材料版本
- Official GPT-5 developer release article; launch-day Wayback replay 2025-08-07 21:34:00 UTC, CDX status 200 text/html; detailed benchmark tables read and transcribed. This is the later of two same-day captures with different FrontierMath nano values.
- 版本公布
- 披露时间
- · 已有首发依据
- 采集时间
- 历史依据
- Wayback CDX capture 20250807213400, status 200 text/html; replay via http://web.archive.org/web/20250807213400id_/https://openai.com/index/introducing-gpt-5-for-developers/ returned HTTP 200. The archived article is dated August 7, 2025.
引用文本
GPT-5 nano · τ²-bench(Telecom section),GPT-5 nano (high)配置,Success rate 35.5%(成绩公布 2025-08-07)。官方出处 Introducing GPT-5 for developers | OpenAI(主文,Official GPT-5 developer release article; launch-day Wayback replay 2025-08-07 21:34:00 UTC, CDX status 200 text/html; detailed benchmark tables read and transcribed. This is the later of two same-day captures with different FrontierMath nano values.)Detailed benchmarks / τ²-bench telecom,https://openai.com/index/introducing-gpt-5-for-developers/。Benchlog https://benchlog.lekshome.cc/evidence/gpt-5-nano/ev-gpt5-nano-tau2-telecom/
关联成绩与条件
| 指标与成绩 | 配置与评测条件 | 证据关系 |
|---|---|---|
| Success rate35.5%原文 35.5% · 单位 % 公布 | GPT-5 nano (high)
| 直接支持该成绩 在排名中查看此成绩 |