Unverified60% confidenceOpinionExact time
OpenAI did not publish official scores for the standard SWBench for GPT-5.2 Codex, suggesting they may not have surpassed Opus 4.5's performance on that benchmark.
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
Related Claims
UnverifiedOpenAI发布材料中未公布Sol在SWE-bench Pro的得分，该项测试目前由Fable 5领跑73% similarUnverifiedFable 5在SWE Bench Pro编程测试上得分80.3分，而OpenAI的GPT 5.5仅为58.6分64% similarPartially VerifiedClaude Opus 4.8在SWE-bench基准测试中得分69.2%，GPT 5.5为58.6%，Google Gemini为54.2%63% similarUnverifiedClaude Opus 4.8在AI Coding基准测试中表现超越GPT-5.562% similarUnverified在Terminal Bench 2.0基准测试中，OpenAI 4.6得分为58%，GPT 5.4得分为75.1%62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/51094API
curl https://kongchang.com/api/v1/knowledge/claims/51094MCP
get_claim(id=51094)