OpenAI Releases GPT-5.3 Instant: Enhancements to ChatGPT, New Features, and Fewer Errors

|
OpenAI Releases GPT-5.3 Instant: Enhancements to ChatGPT, New Features, and Fewer Errors

OpenAI has introduced the updated artificial intelligence model GPT-5.3 Instant for ChatGPT, which features more accurate responses, a more natural communication style, and fewer erroneous replies.

This is reported by Finway

Key Changes in GPT-5.3 Instant

GPT-5.3 Instant has received significant updates aimed at improving everyday communication between users and the language model. Developers focused on three key aspects: tone of communication, relevance of responses, and fluidity of dialogue. The model now better maintains naturalness in conversations, uses categorical phrasing less often, and avoids excessive warnings.

GPT-5.3 Instant is now available for ChatGPT users, as well as for developers through the API named gpt-5.3-chat-latest. Meanwhile, the previous version, GPT-5.2 Instant, will remain available in the Legacy Models section for paid users for another three months and will be retired on June 3, 2026.

“GPT-5.3 Instant is less likely to get stuck in a conversation, reduces excessive cautions, and avoids overly categorical phrasing.”

Fewer Hallucinations and Higher Accuracy of Responses

One of the main advantages of the new update is the reduction in instances where the model either refuses to answer safe questions or begins responses with unnecessary warnings. GPT-5.3 Instant more frequently provides clear and helpful answers without moralizing introductions. At the same time, its handling of web sources has improved: the system better correlates found information with its own knowledge base, allowing it not only to summarize search results but also to explain their context and highlight key points.

According to OpenAI, in tasks involving web search, the number of erroneous responses (“hallucinations”) has decreased by 26.8%, while when working solely with internal knowledge, it has decreased by 19.7%. User feedback ratings indicate a reduction in the frequency of such responses by 22.5% and 9.6%, respectively.

Comparison of responses from GPT 5.2 and GPT 5.3. Data: OpenAI.

The model has also become a more effective tool for writing — it better assists in creating diverse prose, editing materials, and generating content ideas. According to the developers, GPT-5.3 Instant avoids unnecessary phrases and artificial pauses, making the dialogue with users more relaxed and meaningful.

As a reminder, OpenAI previously received $110 billion in investments, and the company’s market valuation reached $730 billion.