Unverified50% confidenceBenchmarkExact time
在 FrontierSWE 基准上，GLM-5.2 仅落后 Opus 4.8 约 1%，以 1% 优势险胜 GPT-5.5，并领先 Opus 4.7 达 11%
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
8/26/2026
First Seen
Valid until: 9/9/2026
Sources
Related Entities
Related Claims
Unverified在 Frontiers-WE 基准上，GLM-5.2 与 Claude Opus 4.8 的差距缩小到不足一分78% similarPartially VerifiedClaude Opus 4.8在SWE-bench基准测试中得分69.2%，GPT 5.5为58.6%，Google Gemini为54.2%77% similarUnverified早期GPT-4在SWE-bench的通过率仅为1.7%，Claude 3 Opus约为4.8%，直到2024年中才突破20%75% similarUnverifiedOpus 4.7在长上下文（50万-100万token）检索方面表现明显退步，远不如DeepSeek和GPT 5.573% similarUnverifiedGLM 5.2在Frontier SW基准测试中得分为74.4%，接近Anthropic Opus 4.8的水平73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/805114API
curl https://kongchang.com/api/v1/knowledge/claims/805114MCP
get_claim(id=805114)