Overview
- OpenAI says GPT-5.3 Instant reduces unnecessary refusals and tones down moralizing preambles before answers.
- The update aims for a more natural style that cuts back on formulaic phrasing and so‑called “cringe” responses.
- Replies are designed to get to the point more quickly with less lead‑up explanation.
- Web integration is updated to deliver richer, better‑contextualized information rather than link lists.
- The rollout starts with the Instant tier, with Thinking and Pro updates promised soon.