Verified75% confidenceBenchmarkExact time
GPT-4 and Claude 3 achieve pass@1 rates exceeding 85% on the HumanEval code benchmark.
3
Sources
75%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
Base44 in Practice: How Claude Code Powered Rapid Scaling from 1 to 80 People
bilibiliGelai_AI7/5/2026
Related Claims
UnverifiedEarly GPT-3's HumanEval Pass@1 was below 10%, while top models such as GPT-4o and Claude 3.5 Sonnet have now consistently exceeded 85%.81% similarUnverifiedThe Pass@1 rate of frontier models such as GPT-4 and Claude on code generation tasks is typically between 60% and 80%, meaning 1 to 2 out of every 5 modifications may have issues.77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/489873API
curl https://kongchang.com/api/v1/knowledge/claims/489873MCP
get_claim(id=489873)