Expired60% confidenceBenchmarkExact time
Since 2024, the scores of various AI programming tools on SWE-bench have surged from initially less than 5% to over 50%.
2
Sources
60%
Confidence
Short-term (~14 days)
Relevance
8/30/2026
First Seen
Valid until: 9/13/2026(expired)
Sources
Related Entities
Related Claims
VerifiedIn early 2024, the strongest AI models scored only in the 10%–20% range on SWE-Bench.77% similarUnverified自2023年GPT-4发布以来，AI在HumanEval等主流编程基准测试上的通过率从不足50%跃升至90%以上74% similarUnverifiedOn SWE-bench, early models had a solve rate below 5%, while mainstream products in 2025 have surpassed 50%.73% similarUnverifiedA 2024 Princeton University study found that content optimized through GEO can increase the probability of being cited by AI by more than 40%.71% similarUnverified新一代模型在SWE-bench等真实软件工程任务上的通过率首次突破50%大关71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/826970API
curl https://kongchang.com/api/v1/knowledge/claims/826970MCP
get_claim(id=826970)