Claude Opus 5vsGPT-5.6 Sol
在 61 个同模式 benchmark 中，Claude Opus 5 整体领先：Claude Opus 5 领先 46 项，GPT-5.6 Sol 领先 14 项，持平 1 项，平均分差 +7.45。另有 7 项测试模式不同，仅供参考。
Claude Opus 5
Anthropic · 2026-07-24 · 推理大模型
GPT-5.6 Sol
OpenAI · 2026-06-26 · 推理大模型
Claude Opus 546 项(75%)持平1(23%)14 项GPT-5.6 Sol
评测分数
按能力类目分组，每组内按分差大小排列；共 61 项。
仓库修复与多文件工程
Claude Opus 5 领先 6/7| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| FrontierSWE v2 | 52.016 / 22Max (With Tools) | 32.209 / 22Max (With Tools) | +19.81 |
| NL2Repo-Bench | 75.301 / 20Max (With Tools) | 56.8011 / 20Max (With Tools) | +18.50 |
| APEX-SWE | 63.703 / 72Max (With Tools) | 4818 / 72Max (With Tools) | +15.70 |
| SWE-Bench Pro V2 Hard | 981 / 12Extra-High (With Tools) | 82.408 / 12Extra-High (With Tools) | +15.60 |
| SWE-Bench Pro V2 | 99.401 / 10Extra-High (With Tools) | 95.506 / 10Extra-High (With Tools) | +3.90 |
| DeepSWE | 73.6512 / 106Max (With Tools) | 72.6718 / 106Max (With Tools) | +0.98 |
| SWE-Atlas-QnA | 52.706 / 9Reported best (effort unspecified) | 53.505 / 9Reported best (effort unspecified) | -0.80 |
数学
GPT-5.6 Sol 领先 3/4| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| ProofBench v1.1 (Lean 4) | 995 / 53Max (With Tools) | 8315 / 53Max (With Tools) | +16 |
| FrontierMath Tier 4 v2 | 73.1719 / 77Max (With Tools) | 82.9313 / 77Max (With Tools) | -9.76 |
| Riemann-bench | 684 / 42Max (With Tools + Internet) | 74.401 / 42Max (With Tools + Internet) | -6.40 |
| FrontierMath v2 | 85.6112 / 120Max (With Tools) | 89.126 / 120Max (With Tools) | -3.51 |
自主开发与终端任务
Claude Opus 5 领先 4/4| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Terminal-Bench 4.0 | 51.8217 / 108Max (With Tools) | 37.2729 / 108Max (With Tools) | +14.55 |
| Terminal-Bench 3.0 | 43.301 / 11Max (With Tools) | 34.402 / 11Max (With Tools) | +8.90 |
| Vibe Code Bench 1-100 | 28.533 / 25Max (With Tools) | 20.0111 / 25Max (With Tools) | +8.52 |
| CursorBench 4.0 | 46.6015 / 75Max (With Tools) | 41.7027 / 75Max (With Tools) | +4.90 |
图像感知与视觉推理
Claude Opus 5 领先 2/3| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| BabyVision | 94.101 / 13Max (With Tools) | 88.906 / 13Max (With Tools) | +5.20 |
| MMMU-Pro | 84.708 / 233Max (No Tools) | 83.4016 / 233Max (No Tools) | +1.30 |
| ZeroBench Main | 522 / 8Max (With Tools) | 531 / 8Max (With Tools) | -1 |
客服与业务流程
Claude Opus 5 领先 2/3| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Public Benefits Bench v1.1 | 76.931 / 50Max (With Tools) | 66.5118 / 50Max (With Tools) | +10.42 |
| τ³-Banking | 48.716 / 170Max (With Tools) | 44.3020 / 170Max (With Tools) | +4.41 |
| SAGE | 49.4321 / 91Max (With Tools) | 52.566 / 91Max (With Tools) | -3.13 |
法律
Claude Opus 5 领先 3/3| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Legal Research Bench | 55.291 / 84Max (With Tools) | 48.0810 / 84Max (With Tools) | +7.21 |
| Harvey Lab-AA | 93.464 / 44Max (With Tools) | 87.1817 / 44Max (With Tools) | +6.28 |
| Harvey's Legal Agent Benchmark | 6.6721 / 51Max (With Tools) | 2.5036 / 51Max (With Tools) | +4.17 |
科学知识与推理
GPT-5.6 Sol 领先 2/3| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| MysteryMechanism | 37.397 / 28Max (With Tools) | 33.3312 / 28Max (With Tools) | +4.06 |
| CritPt | 29.1012 / 208Max (No Tools) | 32.301 / 208Max (No Tools) | -3.20 |
| GPQA Diamond | 87.8865 / 250Thinking Low (No Tools) | 89.9044 / 250Thinking Low (No Tools) | -2.02 |
程序生成与编辑
Claude Opus 5 领先 3/3| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Program Bench | 376 / 18Max (With Tools) | 2311 / 18Max (With Tools) | +14 |
| Vibe Code Bench v1.1 | 88.4010 / 114Max (With Tools) | 80.5026 / 114Max (With Tools) | +7.90 |
| FrontierCode 1.1 Main | 53.404 / 50Reported best (effort unspecified) | 47.5013 / 50Reported best (effort unspecified) | +5.90 |
金融
Claude Opus 5 领先 3/3| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Tax Agent Bench | 75.062 / 71Max (With Tools) | 67.958 / 71Max (With Tools) | +7.11 |
| Finance Agent v2 | 58.6310 / 76Max (With Tools) | 53.7631 / 76Max (With Tools) | +4.87 |
| EMB (Excel Modeling) | 73.566 / 77Max (With Tools) | 72.347 / 77Max (With Tools) | +1.22 |
临床工作与医疗决策
Claude Opus 5 领先 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| MedCode | 63.571 / 107Max (With Tools) | 43.9743 / 107Max (With Tools) | +19.60 |
| MedScribe | 90.984 / 109Max (With Tools) | 85.2331 / 109Max (With Tools) | +5.75 |
多轮记忆与持续上下文
Claude Opus 5 领先 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| EBR-bench | 45.715 / 23Max (With Tools) | 44.766 / 23Max (With Tools) | +0.95 |
| Context Arena | 97.721 / 154Max (No Tools) | 97.632 / 154Max (No Tools) | +0.09 |
抽象归纳与泛化
Claude Opus 5 领先 1/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| ARC-AGI-3 | 30.1610 / 66Thinking High (No Tools) | 2.1522 / 66Thinking High (No Tools) | +28.01 |
| ARC-AGI-1 | 97.5013 / 249Extra-High (No Tools) | 97.5013 / 249Extra-High (No Tools) | 持平 |
文档与图表理解
胶着 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| GDP.pdf | 21.6034 / 126Max (No Tools) | 27.2012 / 126Max (No Tools) | -5.60 |
| Chartography | 845 / 17Max (With Tools) | 79.907 / 17Max (With Tools) | +4.10 |
机器学习工程
Claude Opus 5 领先 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| WeirdML v2 | 91.591 / 52Thinking High (With Tools) | 88.762 / 52Thinking High (With Tools) | +2.83 |
| WeirdML v3 | 0.198 / 24Extra-High (With Tools) | 0.1511 / 24Extra-High (With Tools) | +0.04 |
科学计算与科研编程
胶着 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Terminal-Bench-Science 0.1 | 305 / 18Max (With Tools) | 22.406 / 18Max (With Tools) | +7.60 |
| SciCode | 56.4028 / 137Max (No Tools) | 57.1022 / 137Max (No Tools) | -0.70 |
综合知识考试
胶着 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| HLE | 56.3024 / 244Max (No Tools) | 44.5070 / 244Max (No Tools) | +11.80 |
| HLE-Verified | 54.403 / 6Reported best (effort unspecified) | 54.502 / 6Reported best (effort unspecified) | -0.10 |
跨应用与工具编排
Claude Opus 5 领先 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| APEX-Agents | 65.806 / 61Max (With Tools) | 56.7018 / 61Max (With Tools) | +9.10 |
| Agents' Last Exam | 32.205 / 21Max (With Tools) | 30.608 / 21Max (With Tools) | +1.60 |
跨能力综合测试
胶着 2/2| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| LiveBench 2.0 | 82.098 / 49Max (task-specific tools) | 76.2614 / 49Max (task-specific tools) | +5.83 |
| LiveBench | 80.0812 / 157Max (No Tools) | 81.0510 / 157Max (No Tools) | -0.97 |
专业数据分析
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| AA-AnalystAgent | 53.753 / 29Max (With Tools + Internet) | 47.507 / 29Max (With Tools + Internet) | +6.25 |
写作与创意表达
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Creative Writing | 2,1333 / 114Normal (No Tools) | 1,9729 / 114Normal (No Tools) | +161 |
办公与文档流程
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| AutomationBench | 50.3013 / 148Max (With Tools) | 45.8019 / 148Max (With Tools) | +4.50 |
桌面工作流
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| OSWorld 2.0 | 70.574 / 20Max (With Tools) | 62.6011 / 20Max (With Tools) | +7.97 |
漏洞利用
GPT-5.6 Sol 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| ExploitGym (budget unspecified) | 22.103 / 10Max (With Tools) | 33.702 / 10Max (With Tools) | -11.60 |
生物与基因
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| BioMysteryBench | 79.263 / 21Max (With Tools) | 71.1111 / 21Max (With Tools) | +8.15 |
算法与竞赛编程
GPT-5.6 Sol 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| IOI (Vals v2) | 84.338 / 45Max (With Tools) | 91.175 / 45Max (With Tools) | -6.84 |
维护、优化与工程管理
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Code Migration | 57.478 / 79Max (With Tools) | 52.9212 / 79Max (With Tools) | +4.55 |
跨能力聚合指数
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| Vals Index v2.1 | 63.675 / 38Max (task-specific tools) | 58.019 / 38Max (task-specific tools) | +5.66 |
逻辑与规划
Claude Opus 5 领先 1/1| 评测项 | Claude Opus 5 | GPT-5.6 Sol | 分差 |
|---|---|---|---|
| BALROG | 63.402 / 32Max (With Tools) | 603 / 32Max (With Tools) | +3.40 |
测试模式不同的成绩
共 7 项。两款模型的公开成绩来自不同测试模式，分数不直接可比，不计入胜负和平均分差。
| 评测项 | Claude Opus 5 | GPT-5.6 Sol |
|---|---|---|
| Vals Index | 67.21Max (With Tools) | 72.63Extra-High (No Tools) |
| SimpleBench | 80.60Thinking High (No Tools) | 64.80Extra-High (No Tools) |
| BrowseComp | 90.80Max (With Tools + Internet) | 90.40Reported best (effort unspecified) |
| PostTrain Bench | 35Reported best (effort unspecified) | 36.20Max (With Tools) |
| Text Arena (Coding) | 1,712Max (No Tools) | 1,620Extra-High (No Tools) |
| SWE-Bench Pro - Public | 79.20Max (With Tools) | 64.60Extra-High (With Tools) |
| MCP-Atlas | 85.80Extra-High (With Tools) | 81.80Max (With Tools) |
规格对比
| 字段 | Claude Opus 5 | GPT-5.6 Sol |
|---|---|---|
| 发布机构 | Anthropic | OpenAI |
| 发布时间 | 2026-07-24 | 2026-06-26 |
| 模型类型 | 推理大模型 | 推理大模型 |
| 架构 | 稠密模型 | 稠密模型 |
| 参数规模 | 暂无数据 | 暂无数据 |
| 上下文长度 | 1M | 1.05M |
| 最大输出 | 128K | 128K |
API 调用价格
价格优先使用 DataLearner 配置的 API 记录；缺失项不做推测。
| 价格项 | Claude Opus 5 | GPT-5.6 Sol |
|---|---|---|
| 文本输入 | $5 / 1M tokens | $4 / 1M tokens |
| 文本输出 | $25 / 1M tokens | $20 / 1M tokens |
| 缓存读取 | $0.5 / 1M tokens | $0.4 / 1M tokens |
| 缓存写入 | $6.25 / 1M tokens | $5 / 1M tokens |
小结
- Claude Opus 5在以下类目领先:仓库修复与多文件工程 (6/7)、自主开发与终端任务 (4/4)、图像感知与视觉推理 (2/3)、客服与业务流程 (2/3)、法律 (3/3)、程序生成与编辑 (3/3)、金融 (3/3)、临床工作与医疗决策 (2/2)、多轮记忆与持续上下文 (2/2)、抽象归纳与泛化 (1/2)、机器学习工程 (2/2)、跨应用与工具编排 (2/2)、专业数据分析 (1/1)、写作与创意表达 (1/1)、办公与文档流程 (1/1)、桌面工作流 (1/1)、生物与基因 (1/1)、维护、优化与工程管理 (1/1)、跨能力聚合指数 (1/1)、逻辑与规划 (1/1)
- GPT-5.6 Sol在以下类目领先:数学 (3/4)、科学知识与推理 (2/3)、漏洞利用 (1/1)、算法与竞赛编程 (1/1)
- 胶着类目:文档与图表理解、科学计算与科研编程、综合知识考试、跨能力综合测试
61 个同模式 benchmark 上，Claude Opus 5 平均高出 7.45 分。
单项差距最大的 benchmark：Creative Writing — Claude Opus 5 2,133，GPT-5.6 Sol 1,972（分差 +161）。
本页正文由结构化模型、价格与 benchmark 数据生成，不使用实时 LLM 撰写。