SUIVI QUOTIDIEN DES VERSIONS
OpenAI’s consumer and team AI workspace release notes.
Les explications ci-dessous sont en anglais (aucune version française n’est encore produite).
OpenAI added an optional Advanced Account Security mode for personal accounts. The practical impact is stronger account protection, but recovery becomes stricter, so users must keep recovery keys and passkeys safe.
ChatGPT made model switching less hidden by moving model selection into the composer flow. This reduces friction when choosing between fast, thinking, and pro-grade models before sending a prompt.
OpenAI introduced a faster response path for straightforward information questions. It trades personalization for speed, which is useful for factual lookups but less suitable when memory or prior context matters.
ChatGPT’s image stack moved to Images 2.0 and added a thinking mode for paid users. The important shift is that image generation is becoming more deliberative, not just one-shot rendering.