Unverified50% confidenceOpinionExact time
跑分排名不等于所有场景的实际表现，Terra跑分接近5.5但游戏任务中5.5完成度略高
1
Sources
50%
Confidence
Long-term
Relevance
7/23/2026
First Seen
Sources
GPT-5.6 Three-Model Showdown: Sol vs Terra vs Luna — Which One Should You Pick?
bilibili麦冬AI实验室7/11/2026
Related Claims
UnverifiedFable 5 scored 80.3 on the SWE Bench Pro programming test, while OpenAI's GPT 5.5 scored only 58.6.69% similarUnverifiedThe lowest-purchase-price mechanism was tested on 5–6 numbers, with a success rate of approximately one-third65% similarUnverifiedThe champion team NRes used the Mosaic method (not on Muni's official tool menu), with 4 out of 6 designs succeeding for a hit rate of 66.7%.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/602593API
curl https://kongchang.com/api/v1/knowledge/claims/602593MCP
get_claim(id=602593)