Unverified50% confidenceEventExact time
MiniMax M3在财报分析测试中发现了苹果Q2财报可视化图中的数据错误并主动指出，而Gemini 3.1 Pro将错误原样照抄进表格
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
MiniMax M3实测对比GPT、Claude、Gemini：五个真实任务谁更强
bilibili数码旭6/8/2026
Related Claims
UnverifiedGemini 2.5 Pro在数学能力基准测试中与O3 Mini持平，两者都将其他模型远远甩在身后65% similarUnverified原生Gemini Mac应用据报已进入内部测试阶段64% similarUnverified一位Reddit用户认为Gemini 3.5 Flash在电子表格和文档工作上被严重低估，并称有基准测试证实了这一点64% similarUnverifiedGemini Pro在科研技术路线图生成测试中存在图像模糊、稳定性差、频繁报错无法生成图片的问题63% similarUnverifiedGemini 3在SWE-bench基准测试中智能体场景下的得分远超同类模型63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/49531API
curl https://kongchang.com/api/v1/knowledge/claims/49531MCP
get_claim(id=49531)