Unverified50% confidenceBenchmarkExact time
In EvoArena testing, GPT-5.5, which achieved the highest Step Level score, had an accuracy of just over 50%, while Gemini-3.1 Pro and GLM-5.1, which performed best at the Chain Level, reached only approximately 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
UnverifiedIn the Agents' Last Exam benchmark, GPT-5.6 Sol achieved a score of 53.6, which is 13.1 points higher than Claude Fable 5 using adaptive reasoning.76% similarUnverifiedOn Artificial Analysis's Omniscience hallucination metric, GPT-5.6 Sol Max reached 89%, higher than Claude Fable's 55% and GLM 5.2's 28%.76% similarUnverifiedGPT-5.6 set a new all-time high score of 53.6 on the Agent's Last Exam benchmark, surpassing Claude Opus 4.5 by 13.1 points.75% similarUnverifiedOn the SWE-Bench Pro programming benchmark, Claude Fable 5 leads with a score of 80%, ahead of GPT-5.6 Sol's 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)