Unverified75% confidenceFactTime unknown
Opus 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 5.5
1
Sources
75%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026
Sources
GPT 5.5 vs Claude Code vs DeepSeek V4：三大编码模型实测对比
bilibili量子菠萝_
Related Entities
Related Claims
UnverifiedOpus 4.7每次解题的中位输出token数为60,000，而GPT 5.5仅需16,00079% 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%74% 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%73% similarUnverifiedClaude Opus 4.7比之前版本有进步，但提升不够明显73% similarUnverifiedClaude Opus 5能力接近上一代旗舰Opus水平，但推理成本显著降低73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/2469API
curl https://kongchang.com/api/v1/knowledge/claims/2469MCP
get_claim(id=2469)