OpenAI has released GPT‑5.3 Instant, a new model for ChatGPT that aims to make conversations quicker, more accurate and less stilted.
Reported by PCWorld, the update reduces the tendency for the chatbot to deliver overly cautious or moralising responses, instead providing direct answers with smoother flow.
The company says GPT‑5.3 Instant has achieved measurable improvements in accuracy. Hallucination rates, where the system produces incorrect or misleading information, have dropped by 26.8% in web searches and 19.7% in knowledge base queries. This means users can expect more reliable results when asking for factual information or searching online.
Speed has also been improved. The Instant model is designed to respond faster, cutting down delays that previously made interactions feel sluggish. Context handling has been refined, producing richer and more relevant answers without unnecessary caveats. The update reduces the number of refusals for safe queries, addressing a common frustration where the chatbot sometimes declined to answer straightforward questions.
The rollout is notable because it applies to all ChatGPT users, including those on free accounts. This marks a shift from earlier updates that were often reserved for paying subscribers.
The timing of the update comes as competition in the AI space intensifies. Rivals such as Anthropic and Google are pushing their own conversational models, while OpenAI faces scrutiny over reliability in enterprise settings. GPT‑5.3 Instant is aimed at reassuring both casual users and professional customers that ChatGPT can deliver trustworthy results without unnecessary interruptions.
With GPT‑5.3 Instant now live, ChatGPT is expected to feel less like a lecture and more like a practical tool, offering faster, more accurate and more natural conversations for millions of users worldwide.
The post ChatGPT updates dials down the annoying conversation appeared first on Trusted Reviews.