Unverified50% confidenceBenchmarkExact time
该模型在'未见过环境'测试中达到 76.6% 的导航成功率,比最优单摄像头方案高出 9.7 个百分点
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/16/2026
First Seen
Valid until: 10/14/2026(expired)
Sources
Related Claims
Unverified该模型在最难的未见过环境测试中取得76.6%的导航成功率84% similarUnverified每步60%准确率意味着连续5步后整体成功率约为7.8%(0.6的5次方)66% similarUnverifiedJCE所谓的'可靠'实际只有约60%的准确率,因误差累积连续5步后整体成功率仅剩约7.8%61% similarUnverifiedRoboDojo具身智能测评显示,最强模型得分仅12.8分(人类基准为100分),仿真成功率不足9%,真机成功率12.8%59% similarUnverified移除歧义后所有模型准确率提升26.8至40.2个百分点,Gemini甚至达到81%,表明歧义是深度搜索代理最大性能瓶颈之一58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/536180API
curl https://kongchang.com/api/v1/knowledge/claims/536180MCP
get_claim(id=536180)