Unverified50% confidenceFactExact time
测试的两个项目此前已用Claude Opus 4.5加GPT-5.5的组合写过一轮
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/18/2026
First Seen
Valid until: 10/16/2026
Sources
Related Claims
VerifiedM2.5在编程任务上跑出了接近Claude Opus 4.6和GPT-5.2等旗舰级模型的表现80% similarUnverifiedComposer 2.5在SweepBench多语言测试中与Claude Opus 4.7和GPT 5.5属于同一梯队78% similarVerifiedThePrimeagen在同一天测试了Claude Opus 4.6和GPT 5.3 Codex，用完全相同的任务进行编程对决77% similarUnverifiedGPT-5.6在知识工作、科学与健康领域测试全面领先Claude Opus及Opus 4.576% similarUnverified在多米诺骨牌物理模拟测试中，Opus 4.5使用一次提示即完成五个阵型设计，而GPT-5.2 Codex需要两次提示才达到基本可用状态76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/550732API
curl https://kongchang.com/api/v1/knowledge/claims/550732MCP
get_claim(id=550732)