Model dossier模型档案
GPT-5.5
GPT-5.5 is a high-tier model from OpenAI (Intelligence Score 82.9/100). Strongest in math (97.9) and coding (90.6); best fit: quantitative and scientific work. It trails in preference (45.8). Premium-priced: $5.00 in / $30 out per 1M tokens. Context window 1.1M tokens — long-document friendly.
GPT-5.5 是 OpenAI 的高水平梯队模型(智能评分 82.9/100)。 最强项是数学(97.9分),其次是代码(90.6分);适合数理与科研任务。 短板是偏好(45.8分)。 定价属高端定价:每百万 token 输入 $5.00 / 输出 $30。 上下文 1.1M token,长文档友好。
82.9
Intelligence Score · benchmark coverage 智能评分 · benchmark 覆盖 100%
◈
Benchmark scoresBenchmark 成绩
11 entries 条| Benchmark基准 | Score成绩 | Index指数 | Dated日期 | Measured by测评方 | |
|---|---|---|---|---|---|
| Humanity's Last ExamReasoning | 36.24 | 63 | — | official官方榜 | source ↗×6⚠±15.96 |
| ↳ ⚡high | 43.6 | — | — | 3rd-party第三方 | source ↗ |
| GPQA DiamondReasoning | 93.5⚡high | 98 | — | 3rd-party第三方 | source ↗×7 |
| ↳ ⚡max | 93.6 | — | — | vendor-reported厂商自报 | source ↗ |
| MMLU-ProKnowledge | 86.1 | 91 | — | 3rd-party第三方 | source ↗ |
| SWE-bench VerifiedCoding | 87 | 90 | 2026-08 | 3rd-party第三方 | source ↗×7⚠±6.1 |
| AIME 2025Math | 100 | 100 | — | 3rd-party第三方 | source ↗×5⚠±4.8 |
| LiveCodeBenchCoding | 85.3 | 91 | — | 3rd-party第三方 | source ↗×2⚠±17.9 |
| FrontierMathMath | 85⚡max | 96 | — | official官方榜 | source ↗×5⚠±33.6 |
| Terminal-BenchCoding | 82.7 | 90 | — | official官方榜 | source ↗×4 |
| ↳ ⚡off | 83.37 | — | — | 3rd-party第三方 | source ↗ |
| τ²-benchAgent | 46.39 | 47 | 2026-03-11 | official官方榜 | source ↗×5⚠±51.61 |
| LMArena (Chatbot Arena)Preference | 1482⚡high | 46 | 2026-08-12 | official官方榜 | source ↗×4 |
| WebArenaAgent | 59 | 64 | — | 3rd-party第三方 | source ↗×2⚠±3.7 |
≡
Specs & pricing规格与价格
| Provider厂商 | OpenAI |
|---|---|
| Released发布 | 2026-04-24 |
| Context window上下文 | 1.1M tokens |
| Max output最大输出 | 128K tokens |
| Modality模态 | text+image+file->text |
| Reasoning tiers推理档位 | ⚡low ⚡medium ⚡high ⚡max default默认 medium · minimal / low / medium / high / xhigh |
| API input price输入价格 | $5.00 / 1M tokens |
| API output price输出价格 | $30 / 1M tokens |
⛓
Tools using this model使用它的工具
⚔
Head-to-head正面对比
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。