
OpenAI has released GPT-5.3 Instant, an update to ChatGPT aimed at reducing what users described as a 'cringe' tone characterized by overly cautious or preachy disclaimers. The new model focuses on improving conversational flow, relevance, and accuracy, cutting unnecessary caveats and refusals. It also reduces hallucination rates and balances web-sourced information with internal knowledge. The update arrives amid criticism of OpenAI's Pentagon agreement, though the company emphasizes enhanced user experience and response quality.
Select a news story to see related coverage from other media outlets.