Unverified50% confidenceBenchmarkExact time
Under the same concise prompt and reference-overlap scorer, the published 406M specialized model leads at least five commercially hosted large models by at least 6.2 points on ROUGE-1.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/23/2026
First Seen
Valid until: 12/22/2026
Sources
Related Entities
Related Claims
UnverifiedThe fine-tuned 406M model achieved 36.33 ROUGE-1 on the test set, while the authors' own 1.2B system achieved 35.41 ROUGE-1, and the two cannot be statistically distinguished.69% similarUnverified在MoE模型中每个token只激活少量Expert（通常2-4个，Top-K路由），实际计算量可能仅相当于6-7B的Dense模型64% similarUnverifiedGLM4 scored 46.2% on the SWE-Bench benchmark, significantly outperforming the previous generation model.64% similarUnverifiedGLM-5.2 topped the leaderboard in DesignArena with a score of 1360.64% similarUnverifiedGLM-5.3在六项评测中的成绩为：终端编程基准3.0得28.3、深度软件工程基准得66.9、智能体中场考试得28.5、自动化基准得48.2、带工具的人类终极考试得62.5、GDP价值评估AA2版得17.6963% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/942880API
curl https://kongchang.com/api/v1/knowledge/claims/942880MCP
get_claim(id=942880)