Expired85% confidenceFactTime unknown
The average change size of PRs submitted by agents is significantly smaller than those of humans, mostly small changes of a few dozen lines, whereas human developers' PRs frequently contain functional submissions of hundreds or even thousands of lines.
1
Sources
85%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
Valid until: 8/29/2026(expired)
Sources
Related Claims
UnverifiedThe issue resolution rate of agent PRs is noticeably lower than that of humans; although the PRs can pass review, they do not fully resolve the corresponding requirements.65% similarUnverifiedIn large open-source projects, the PR merge rate is typically far lower than the submission rate.64% similarUnverifiedOn cognitive tasks, humans can achieve approximately 80%, while SOTA multimodal models achieve only 30% to 50%, sometimes approaching random guessing63% similarUnverifiedThe top 1% of developers produce 46 times the daily AI coding output of the median developer and merge 15 times as many PRs per week as ordinary developers.62% similarUnverifiedUsing the number of PRs and code revision counts to measure development efficiency is controversial, as there is no positive correlation between lines of code or PR counts and software quality.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/3078API
curl https://kongchang.com/api/v1/knowledge/claims/3078MCP
get_claim(id=3078)