Expired85% confidenceFactTime unknown
Gemini 2.5 Pro scored 18.8% on the Humanities Last Exam benchmark, far exceeding O3 Mini's 14% and other models' scores below 10%.
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
6/1/2026
First Seen
Valid until: 8/30/2026(expired)
Sources
Gemini 2.5 Pro编程实测：Rust重构吊打Claude，前端却翻车
bilibili程序员韩老魔
Related Entities
Related Claims
UnverifiedGemini 3.5 Flash scored 40.2% on Humanities Last Exam, falling behind Opus's 46.9% and Pro's 44.4%.83% similarUnverifiedGemini 3 significantly outperforms comparable models on the SWE-bench benchmark in agentic scenarios.78% similarUnverifiedIn the GDPo expert task evaluation, Gemini 3.1 Pro scored 1317, below the equivalent level of Anthropic Claude.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/20627API
curl https://kongchang.com/api/v1/knowledge/claims/20627MCP
get_claim(id=20627)