OpenAI’s new GPT-5.3 Instant version should provide fewer excessively long answers.
OpenAI has rolled out GPT-5.3 Instant as the new standard experience in ChatGPT. This should result in smoother, more natural conversations and fewer excessive safety disclaimers.
Less wordy
According to OpenAI, the update focuses on the daily user experience: tone, relevance, and conversational flow. Previous models, such as GPT-5.2 Instant, were criticized for sometimes unnecessarily refusing questions or providing long safety explanations before offering an answer. With GPT-5.3 Instant, ChatGPT should get to the point faster and react less over-cautiously.
OpenAI states that the new version is also more reliable. Internal tests show a reduction in hallucinations by about 27 percent when the model uses web information and about 20 percent when it relies solely on training data. Additionally, the model received improvements in writing style and creativity, for example, when writing stories or poetry.
GPT-5.2 Instant remains available
GPT-5.3 Instant replaces GPT-5.2 Instant as the standard model in ChatGPT, although the older version remains available to paying users via legacy options.
The update led to mixed reactions. Some users appreciate the more direct answers, while others criticize OpenAI’s recent business decisions or remain nostalgic for older models like GPT-4o.