Head-to-head正面对比
Claude Opus 5 vs GPT-5.6 Sol
Anthropic meets OpenAI: sourced benchmark scores side by side, dimension by dimension, with prices and specs. Every conclusion below is derived from the published dataset — no opinions, no sponsorship.
Anthropic 对 OpenAI:带来源的 benchmark 成绩逐项并排,附价格与规格。下面每条结论都从公开数据集推导——没有主观意见,没有厂商赞助。
◈
The verdict, derived from data结论(全部由数据推导)
- GPT-5.6 Sol leads overall: 93.4 vs 88 on the Intelligence Score.总分上 GPT-5.6 Sol 领先:93.4 对 88。
- Claude Opus 5 is stronger in reasoning, preference.Claude Opus 5 在推理、偏好上更强。
- GPT-5.6 Sol is stronger in coding, math, agent.GPT-5.6 Sol 在代码、数学、智能体上更强。
- GPT-5.6 Sol is 2.5× cheaper on blended price ($6.00 vs $15 per 1M tokens, in/out average).混合价(输入输出均值)上 GPT-5.6 Sol 便宜 2.5 倍:$6.00 对 $15/百万 token。
● Claude Opus 5 ● GPT-5.6 Sol · missing dimensions count as 0 in the chart缺数据的维度在图上按 0 计
≡
Spec by spec逐项规格
| Claude Opus 5 | GPT-5.6 Sol | |
|---|---|---|
| Provider厂商 | Anthropic | OpenAI |
| Released发布 | 2026-07-24 | 2026-07-09 |
| Intelligence Score智能评分 | 88 | 93.4 |
| Benchmark coveragebenchmark 覆盖 | 100% | 100% |
| Reasoning推理 | 96.4 | 90.4 |
| Coding代码 | 85.9 | 96.3 |
| Knowledge知识 | 96.4 | 93.8 |
| Math数学 | 68.1 | 98.4 |
| Agent智能体 | 78.6 | 92.9 |
| Preference偏好 | 100 | 81 |
| API input $/1M输入价 $/百万 | $5.00 | $2.00 |
| API output $/1M输出价 $/百万 | $25 | $10 |
| Blended $/1M混合价 $/百万 | $15 | $6.00 |
| Value (score per $)性价比(分数/美元) | 5.9 | 15.6 |
| Context window上下文 | 1M | 1.1M |
| Max output最大输出 | 128K | 128K |
| Reasoning tiers推理档位 | low / medium / high / x-high | none / minimal / low / medium / high / xhigh |
| Open weights开放权重 | No否 | No否 |
| CN-direct国内直连 | No否 | No否 |
| Free tier免费层 | No否 | No否 |
⚔
Benchmark by benchmark逐项 benchmark
| Benchmark基准 | Claude Opus 5 | GPT-5.6 Sol |
|---|---|---|
| Humanity's Last Exam | 54.4⚡max | 47.2⚡high |
| GPQA Diamond | 93.7 | 94.1⚡max |
| MMLU-Pro | 91.59 | 89.1 |
| SWE-bench Verified | 96 | 96.2⚡max |
| AIME 2025 | 97 | 97 |
| LiveCodeBench | 89.03 | 81 |
| FrontierMath | 31.25 | 89 |
| Terminal-Bench | 43.5⚡max | 91.9 |
| τ²-bench | 82.78 | 85.1 |
| LMArena (Chatbot Arena) | 1699⚡max | 1623⚡max |
| WebArena | 68 | 92.2 |
Claude Opus 5 dossier → 档案 → · GPT-5.6 Sol dossier → 档案 → · How we score评分方法