Unverified50% confidenceFactExact time
Claude Opus 4.8在真实工程项目Bug修复任务中获得8分，是该任务中得分最高的模型
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
5大模型Coding实战横评：Claude、GPT、DeepSeek、M3谁是工程利器
bilibili不正经的前端啊6/8/2026
Related Claims
UnverifiedClaude Opus 4.8 claimed completion after 8 minutes on its first attempt on the same Cal.com bug fix task but did not actually fix the bug, only succeeding after thinking intensity was set to high73% similarUnverifiedClaude Opus 4.6在SWE-bench基准测试中在现实世界的专家级任务中表现领先72% similarUnverifiedClaude Opus 4.8 has reached top-tier performance among its generation across coding, agentic tasks, reasoning, and knowledge work according to Anthropic71% similarUnverifiedClaude Opus在测试中7分30秒一次成功完成任务，动画流畅、功能完整，排名第一68% similarUnverifiedClaude Opus 4.8在智能体实战跑分中拿下1890分，比上一代高出137分，甩开GPT 5.5达121分68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/48727API
curl https://kongchang.com/api/v1/knowledge/claims/48727MCP
get_claim(id=48727)