Unverified50% confidenceBenchmarkExact time
在285次实测运行中，Gemini取得283次语义通过、2次语义部分通过、0次可复现语义失败
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/23/2026
First Seen
Valid until: 11/21/2026
Sources
Related Entities
Related Claims
UnverifiedThe volume of code generated by Claude Sonnet 4.6 is more than twice that of Gemini 3.1 Pro High (627,000 lines vs. 307,000 lines).60% similarUnverifiedLLaMA 2的32000词表在4096维隐层下嵌入层需占用约1.3亿参数59% similarUnverifiedIn the SaaS dashboard test, GLM 4.7 consumed only 23,000 tokens, far fewer than Gemini and Claude.56% similarVerifiedGemini 3.1 Pro achieved 77% on the advanced reasoning benchmark, while Claude Opus 4.6 scored 68%.55% similarUnverifiedDiscoBench最终生成211个测试样本，包含463个歧义实例55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/791468API
curl https://kongchang.com/api/v1/knowledge/claims/791468MCP
get_claim(id=791468)