Model dossier模型档案
gpt-oss-120b
gpt-oss-120b is a mid-tier model from OpenAI (Intelligence Score 65/100). Strongest in math (93.4) and knowledge (85.3); best fit: quantitative and scientific work. It trails in preference (35.5). Budget-priced: $0.150 in / $0.600 out per 1M tokens. Open weights — self-hostable.
gpt-oss-120b 是 OpenAI 的中端梯队模型(智能评分 65/100)。 最强项是数学(93.4分),其次是知识(85.3分);适合数理与科研任务。 短板是偏好(35.5分)。 定价属低价档:每百万 token 输入 $0.150 / 输出 $0.600。 开放权重,可自托管。
open weights开放权重
65
Intelligence Score · benchmark coverage 智能评分 · benchmark 覆盖 88%
◈
Benchmark scoresBenchmark 成绩
9 entries 条| Benchmark基准 | Score成绩 | Index指数 | Dated日期 | Measured by测评方 | |
|---|---|---|---|---|---|
| Humanity's Last ExamReasoning | 19.6 | 34 | — | official官方榜 | source ↗×3⚠±4.7 |
| GPQA DiamondReasoning | 80.1⚡high | 84 | — | 3rd-party第三方 | source ↗×3 |
| MMLU-ProKnowledge | 79 | 85 | — | 3rd-party第三方 | source ↗×3⚠±11 |
| SWE-bench VerifiedCoding | 60.7 | 63 | 2026-08 | 3rd-party第三方 | source ↗×3 |
| AIME 2025Math | 93.4 | 93 | — | 3rd-party第三方 | source ↗×3⚠±4.5 |
| ↳ ⚡high | 92.5 | — | — | vendor-reported厂商自报 | source ↗ |
| LiveCodeBenchCoding | 70.7 | 76 | — | 3rd-party第三方 | source ↗×3⚠±13.98 |
| Terminal-BenchCoding | 18.7 | 20 | — | 3rd-party第三方 | source ↗ |
| τ²-benchAgent | 65.8⚡high | 67 | 2026-06-10 | 3rd-party第三方 | source ↗×2⚠±20.8 |
| ↳ ⚡low | 45 | — | 2026-06-10 | 3rd-party第三方 | source ↗ |
| LMArena (Chatbot Arena)Preference | 1365 | 36 | — | 3rd-party第三方 | source ↗×2 |
≡
Specs & pricing规格与价格
| Provider厂商 | OpenAI |
|---|---|
| Released发布 | 2025-08-05 |
| Context window上下文 | 131K tokens |
| Max output最大输出 | 66K tokens |
| Modality模态 | text->text |
| Reasoning tiers推理档位 | not documented未见公开文档 |
| API input price输入价格 | $0.150 / 1M tokens |
| API output price输出价格 | $0.600 / 1M tokens |
⛓
Tools using this model使用它的工具
How we score →评分方法 → · Catalog & pricing via the official OpenRouter API.模型目录与价格来自 OpenRouter 官方 API。