Unverified85% confidenceFactTime unknown

Gemini 2.5 Pro在Humanities Last Exam基准测试中得分18.8%，远超O3 Mini的14%和其他模型的10%以下

1

Sources

85%

Confidence

Medium-term (~90 days)

Relevance

6/1/2026

First Seen

Valid until: 8/30/2026

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

Gemini 2.5 Promodel O3 Minimodel Humanities Last Examconcept

Related Claims

UnverifiedGemini 3.5 Flash在Humanities Last Exam中得分40.2%，落后于Opus的46.9%和Pro的44.4%83% similar UnverifiedGemini 3在SWE-bench基准测试中智能体场景下的得分远超同类模型78% similar Unverified在GDPo专家任务评估中，Gemini 3.1 Pro得分1317，低于Anthropic Claude的同等水平76% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/20627

API

curl https://kongchang.com/api/v1/knowledge/claims/20627

MCP

get_claim(id=20627)