Unverified50% confidenceBenchmarkExact time
在EvoArena测试中，Step Level最高的GPT-5.5只有50%出头的准确率，Chain Level表现最好的Gemini-3.1 Pro和GLM-5.1也只有约28%-29%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/16/2026
First Seen
Valid until: 12/15/2026
Sources
Related Entities
Related Claims
Unverified在Agents' Last Exam评测中，GPT-5.6 Sol取得53.6分，比采用自适应推理的Claude Fable 5高出13.1分76% similarUnverified在Artificial Analysis的Omniscience幻觉指标上，GPT-5.6 Sol Max高达89%，高于Claude Fable的55%和GLM 5.2的28%76% similarUnverifiedGPT-5.6在Agent's Last Exam测试中以53.6分创下历史新高，领先Claude Opus 4.5达13.1分75% similarUnverified在SWE-Bench Pro编程评测上，Claude Fable 5以80%的成绩领先GPT-5.6 Sol的64.6%74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/924244API
curl https://kongchang.com/api/v1/knowledge/claims/924244MCP
get_claim(id=924244)