Unverified50% confidenceBenchmarkExact time
MiniMax M3 的基准测试成绩亮眼，部分指标据称可比肩 GPT-5.5
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/10/2026
First Seen
Valid until: 10/8/2026
Sources
实测：GPT-5.4 mini解决国产模型搞不定的崩溃Bug
bilibili库洛米大人的大人6/25/2026
Related Claims
UnverifiedMiniMax M2.1在SWE-Bench基准测试上的表现持平甚至超过Claude Sonnet 4.576% similarUnverifiedMiniMax M2.7在SWE-Pro基准测试上得分56.22%74% similarUnverified在Artificial Analysis第三方评测的编码指数（推理模式）上，GPT-5 Mini High的得分比Haiku 4.5高出9个点71% similarUnverifiedMiniMax M2.7在VibePro基准测试上得分55.6%71% similarUnverifiedMiniMax M3 在拖拽看板测试中得分60分，拖拽逻辑完整但缺乏国际化支持和响应式设计70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/465664API
curl https://kongchang.com/api/v1/knowledge/claims/465664MCP
get_claim(id=465664)