Krux
March 5, 2026
ChatGPT's Most-Used Model Just Stopped Being Annoying
Published: March 5, 2026 at 12:42 AM
Updated: March 5, 2026 at 12:42 AM
100-word summary
OpenAI shipped GPT-5.3 Instant, promising fewer refusals and less "cringe" in everyday conversations. The update to ChatGPT's workhorse model is live for all users today. Instead of hedging every answer with caveats, it now delivers more direct responses and pulls better context from web searches. The change targets the friction that makes ChatGPT feel like talking to an over-cautious assistant. OpenAI kept the old GPT-5.2 Instant around for three months as a safety net before retiring it June 3. Non-English performance still wobbles, but the shift shows OpenAI finally tuning for how people actually use the thing daily.
What happened
OpenAI shipped GPT-5.3 Instant, promising fewer refusals and less "cringe" in everyday conversations. The update to ChatGPT's workhorse model is live for all users today. Instead of hedging every answer with caveats, it now delivers more direct responses and pulls better context from web searches. The change targets the friction that makes ChatGPT feel like talking to an over-cautious assistant. OpenAI kept the old GPT-5.2 Instant around for three months as a safety net before retiring it June 3.
Why it matters
Non-English performance still wobbles, but the shift shows OpenAI finally tuning for how people actually use the thing daily.