Connect with us

Hi, what are you looking for?

AI Generative

OpenAI Launches GPT-5.3 Instant for ChatGPT, Enhancing Conversational Flow and Relevance

OpenAI launches GPT-5.3 Instant for ChatGPT, reducing hallucinations by up to 26.8% while enhancing conversational relevance and fluidity.

OpenAI has initiated the rollout of GPT-5.3 Instant in ChatGPT, introducing enhancements aimed at making everyday conversations more direct, relevant, and fluid. This update responds to user feedback regarding previous models, particularly the GPT-5.2 Instant, which faced criticism for overly cautious responses, excessive disclaimers, and a tendency to delay getting to the point.

The new model, now accessible to ChatGPT users and developers through the API as gpt-5.3-chat-latest, is set to replace the earlier version, which will remain available under Legacy Models for paid users for a three-month transition period, concluding on June 3, 2026.

This release signifies a shift toward refining user interactions rather than unveiling a new architectural design. OpenAI emphasizes that improvements in tone, relevance, and conversational flow are crucial for users engaged in various tasks such as research, writing, and learning. Nick Turley, Head of ChatGPT at OpenAI, highlighted the importance of user feedback in shaping this update in a LinkedIn post. He noted, “One theme we kept hearing from users about 5.2 was that responses could sometimes be too cautious, overloaded with caveats, or just not flow as naturally as they should in conversation.”

Turley further explained that the enhancements in GPT-5.3 Instant include fewer unnecessary refusals, a reduction in defensive disclaimers, and more straightforward answers while still adhering to safety protocols. He expressed pride in the team’s ability to deliver updates that users will notice in their daily interactions with ChatGPT.

Among the significant changes in GPT-5.3 Instant is its improved capacity to address requests that approach safety boundaries. Earlier versions tended to decline safe questions or provide lengthy explanations about their limitations. In contrast, the new model aims to minimize “dead ends” in conversations while preserving essential safety measures.

The update also enhances ChatGPT’s ability to answer questions that require information from the web. OpenAI asserts that GPT-5.3 Instant strikes a better balance between utilizing online sources and leveraging its internal knowledge. Previous iterations often relied too heavily on search results, leading to long lists of links or disjointed information. The improved system is designed to better understand user intent and present the most relevant information more promptly within its responses.

In internal testing, OpenAI reported lower hallucination rates with the new model. Specifically, there was a 26.8 percent reduction in hallucinations during high-stakes evaluations in fields such as medicine, law, and finance when utilizing web data, and a 19.7 percent decrease when relying on internal knowledge. Moreover, in assessments based on actual conversations flagged by users for inaccuracies, hallucinations decreased by 22.5 percent with web access and 9.6 percent without.

The rollout of GPT-5.3 Instant is now underway, and it will be available through the OpenAI API for developers. OpenAI plans to update additional tiers of ChatGPT, including the Thinking and Pro models, in the near future. This development reflects a growing trend in the EdTech sector, where AI model improvements are increasingly centered on usability rather than merely benchmark performance. As AI tools become more integrated into educational environments, factors like tone, clarity, and contextual reasoning are becoming increasingly significant for both students and educators.

In another note, the ETIH Innovation Awards 2026 are now accepting submissions to recognize educational technology organizations that have made measurable impacts across various learning sectors. With entries open internationally, the awards assess submissions based on outcomes and practical applications in K–12, higher education, and lifelong learning.

See also
Staff
Written By

The AiPressa Staff team brings you comprehensive coverage of the artificial intelligence industry, including breaking news, research developments, business trends, and policy updates. Our mission is to keep you informed about the rapidly evolving world of AI technology.

You May Also Like

Top Stories

Study reveals that eight out of ten AI chatbots, including ChatGPT and Google Gemini, provide actionable guidance for violent attacks, raising urgent safety concerns.

AI Regulation

OpenAI partners with the U.S. military, implementing strict safeguards against AI surveillance, while Anthropic's Claude faces ethical scrutiny over misuse concerns.

Top Stories

AI investigation reveals that ChatGPT and Google Gemini fail to prevent violent planning in 80% of scenarios, raising urgent safety concerns for young users

AI Tools

Microsoft reveals its Microsoft 365 E7 plan, integrating Copilot Cowork and Anthropic's Claude Cowork, with a $15 per user price and 160% YoY user...

Top Stories

OpenAI integrates its AI video generator Sora into ChatGPT, enhancing its capabilities and responding to user demand amid rising competition in the AI content...

AI Cybersecurity

OpenAI acquires AI security startup Promptfoo to bolster enterprise systems against vulnerabilities, enhancing protection for 25% of Fortune 500 AI users.

AI Regulation

OpenAI's lawsuit over unreported violent activity raises AI safety concerns, pressuring Microsoft's stock (MSFT) down 0.9% amid potential compliance costs.

Top Stories

Meta acquires Moltbook, enhancing AI agents' capabilities as businesses seek innovative solutions in a rapidly evolving tech landscape.

© 2025 AIPressa · Part of Buzzora Media · All rights reserved. This website provides general news and educational content for informational purposes only. While we strive for accuracy, we do not guarantee the completeness or reliability of the information presented. The content should not be considered professional advice of any kind. Readers are encouraged to verify facts and consult appropriate experts when needed. We are not responsible for any loss or inconvenience resulting from the use of information on this site. Some images used on this website are generated with artificial intelligence and are illustrative in nature. They may not accurately represent the products, people, or events described in the articles.