Unverified50% confidenceBenchmarkExact time
When checking code against written specifications, GPT-4o falsely failed 26.2% and 35.9% of correct programs on two benchmarks respectively, and if it was also required to explain and fix the code, the false-failure rate soared to 73.2% and 87.9%.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedA 2023 study found that GPT-4 produced errors in approximately 15–20% of cases when generating code documentation.80% similarUnverifiedGPT-5.4及更早版本在提取文档中的数字时频繁出错，包括小数点位置偏移、千分位分隔符误读、负号遗漏等问题76% similarUnverifiedGPT-5.4 was reversed by a "trusted source" on 44.7% of questions, Grok-4.20 had a reversal rate as high as 87.5%, while Gemini-3.1-Pro was nearly unaffected at just 0.6%.75% similarUnverifiedGPT-6's error response rate has dropped to 4.1% and its false claim rate to 0.9%, an improvement over GPT-5.6's 6.3% and 1.8% respectively.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1026032API
curl https://kongchang.com/api/v1/knowledge/claims/1026032MCP
get_claim(id=1026032)