Expired50% confidenceBenchmarkExact time
In the Agent Test: Last Exam evaluation, Soul scored 53.6 points and Anthropic's Opus 4.5 scored 40.5 points.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/11/2026
First Seen
Valid until: 10/9/2026(expired)
Sources
GPT-5.6发布：三款新模型上线，Codex正式并入ChatGPT
bilibiliAI学长小林7/10/2026
Related Claims
UnverifiedOn the Agent's Last Exam benchmark, GPT-5.6 Soul (extra high tier) scored nearly 54%, while Fable (max tier) scored 45%.82% similarUnverifiedClaude Sonnet 4.6 improved from 17.7% to 33% on the Humanity's Last Exam (no tools) test.61% similarUnverifiedSol scored 53.6 on Agent's Last Exam, which is 13.1 points higher than Claude Fable 5.60% similarUnverifiedDeepSeek V4 Pro scored 37.7 on the Humanity's Last Exam benchmark.60% 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.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/486691API
curl https://kongchang.com/api/v1/knowledge/claims/486691MCP
get_claim(id=486691)