The AI tools index that doesn't waste your time.不浪费你时间的 AI 工具索引。
Head-to-head正面对比

Claude Sonnet 5 vs GPT-5.5

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)

ReaCodKnoMatAgtPre

Claude Sonnet 5   GPT-5.5 · 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 Sonnet 5GPT-5.5
Provider厂商AnthropicOpenAI
Released发布2026-06-302026-04-24
Intelligence Score智能评分83.682.9
Benchmark coveragebenchmark 覆盖100%100%
Reasoning推理89.180.5
Coding代码88.290.6
Knowledge知识92.290.6
Math数学76.297.9
Agent智能体75.555.4
Preference偏好40.845.8
API input $/1M输入价 $/百万$2.00$5.00
API output $/1M输出价 $/百万$10$30
Blended $/1M混合价 $/百万$6.00$18
Value (score per $)性价比(分数/美元)13.94.7
Context window上下文1M1.1M
Max output最大输出128K128K
Reasoning tiers推理档位low / medium / highminimal / low / medium / high / xhigh
Open weights开放权重NoNo
CN-direct国内直连NoNo
Free tier免费层NoNo

Benchmark by benchmark逐项 benchmark

Benchmark基准Claude Sonnet 5GPT-5.5
Humanity's Last Exam57.436.24
GPQA Diamond74.693.5⚡high
MMLU-Pro87.5586.1
SWE-bench Verified85.287
AIME 202557.4100
LiveCodeBench82.4385.3
FrontierMath8785⚡max
Terminal-Bench80.482.7
τ²-bench80.446.39
LMArena (Chatbot Arena)1462⚡high1482⚡high
WebArena64.559

Claude Sonnet 5 dossier → 档案 → · GPT-5.5 dossier → 档案 → · How we score评分方法