Unverified50% confidenceBenchmarkExact time
在Agent Test: Last Exam测试中，Soul得分53.6分，Anthropic的Opus 4.5得分40.5分
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/11/2026
First Seen
Valid until: 10/9/2026
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
Unverified在Agent's Last Exam基准测试中，GPT-5.6 Soul（extra high档）得分近54%，Fable（max档）为45%82% similarUnverifiedClaude Sonnet 4.6在Humanity's Last Exam（无工具）测试中从17.7%提升至33%61% similarUnverifiedSol 在 Agent's Last Exam 中取得 53.6 分，比 Claude Fable 5 高出 13.1 分60% similarUnverifiedDeepSeek V4 Pro在Humanity's Last Exam基准测试中得分37.760% 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)