[KongchangAI]
Unverified50% confidenceFactExact time

Gemini 3.1 Pro High在SWE-bench测试中通过率为84.17%，排名第一，生成代码量为307,000行，Bug密度为614个/百万行代码，安全问题为210个/百万行代码

1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026

Sources

Sonar实测53个大模型代码质量：Claude安全漏洞最多，GPT-5代码量暴增5倍
bilibiliAI_Express6/1/2026

Related Claims

UnverifiedGemini 3.1 Pro High generated 307,000 lines of code with a bug density of 614 per million lines and security issues of 210 per million lines84% similarUnverifiedGemini 3.5 Flash的编码能力基准测试得分为76.2%78% similar

Cite This Claim

Stable URI
https://kongchang.com/claim/51216
API
curl https://kongchang.com/api/v1/knowledge/claims/51216
MCP
get_claim(id=51216)