Unverified60% confidenceFactExact time
When OpenAI's API endpoint names remain unchanged, the underlying model weights may be silently updated; a 2023 joint study by Stanford and UC Berkeley found significant performance differences in GPT-4's responses to the same prompts across different time periods, with accuracy on certain tasks dropping sharply from 97% to 2%.
2
Sources
60%
Confidence
Long-term
Relevance
8/12/2026
First Seen
Sources
Related Claims
VerifiedA 2023 Stanford University study found that GPT-4's performance on certain tasks showed a measurable decline over several months.79% similarUnverifiedA 2023 Stanford University study found that GPT-4's accuracy on open-domain question-answering tasks was approximately 40 percentage points higher than that of traditional NLU systems.79% similarVerifiedStanford University once published a research report documenting performance fluctuations of GPT-4 on specific tasks across different time periods.77% similarUnverified2023 年曾有多起 OpenAI API 用户报告 GPT-4 Turbo 的 Token 计数与预期不符，最终确认是系统提示词被重复计入的 bug76% similarUnverifiedOpenAI's ChatGPT Plus has usage frequency limits for GPT-4o, and exceeding them results in a downgrade to a lighter model.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/736837API
curl https://kongchang.com/api/v1/knowledge/claims/736837MCP
get_claim(id=736837)