OpenAI API Content Controls Reported Stricter and Buggier Than ChatGPT

MParakhin · x · 2026-08-02

Mikhail Parakhin points out that OpenAI's API content controls are much stricter—and arguably sloppier—compared to ChatGPT or Codex itself. He shared a case where he helped hedge fund buddies debug an intermittent API error, discovering that simply asking about corporate actions of biology-related companies caused the API to randomly barf and reject the request.

Original post →

More from Models

Models channel →