Microsoft Integrates OpenAI’s New GPT-5.3 Instant into 365 Copilot.

Last night, OpenAI announced GPT-5.3 Instant—the lightning-fast successor to its most widely used ChatGPT model, GPT-5.2 Instant. The new model is rolling out immediately to Microsoft 365 Copilot, with API access via Microsoft Foundry slated to arrive very soon.

[ 💰 Ad Slot / Рекламный блок ]

According to OpenAI, this latest iteration delivers more accurate, contextually rich web results. Rather than simply chasing abstract AI benchmark scores, the focus has shifted entirely toward improving tone, relevance, and conversational fluency. GPT-5.3 Instant is now available to all ChatGPT users, officially replacing GPT-5.2. However, paid subscribers will retain access to GPT-5.2 Instant for three months via the "Legacy Models" selector.

The tone of GPT-5.2 could sometimes feel 'clunky,' coming across as overly pushy or making baseless assumptions about a user's intent or emotions," OpenAI stated. "This update introduces a more focused yet natural conversational style, reducing unnecessary declarative phrases and filler like 'Stop. Take a breath.' We are also working to ensure ChatGPT's personality remains consistent across different conversations and updates.

Following suit, Microsoft announced that GPT-5.3 Instant is now live for users of Microsoft 365 Copilot and Copilot Studio. In Microsoft 365 Copilot Chat, the new model appears as "GPT-5.3 Quick response" in the model selector, while in Copilot Studio, it is labeled "GPT-5.3 Chat." Microsoft noted that paid Copilot users will receive priority access to the new model, while free users will be placed on standard access tiers. The model is also making its way to Microsoft Foundry via API, empowering developers and enterprises to build custom AI-driven applications.

Microsoft emphasized several key upgrades that make GPT-5.3 Instant particularly suited for enterprise environments:

• Proportional Safety: Reduces frustrating "dead ends" with fewer unnecessary refusals, providing more balanced and secure responses.
• Adaptive Steering: When a prompt requires adjustment, the model reframes it to meet constraints while keeping the conversation flowing smoothly.
• Grounded Resolution: Integrated web search combined with the model’s reasoning capabilities delivers actionable, well-founded answers rather than a loose list of hyperlinks. This enables end-to-end solutions for IT support and policy workflows while maintaining strict enterprise control and traceability.
• Reliable Performance: At scale, organizations get more consistent outputs—better tone, clearer explanations, stronger instruction-following, and less need for constant prompt engineering.

OpenAI also confirmed that updates to its flagship Thinking and Pro models are coming soon. These upgraded models will similarly be integrated across Microsoft's product ecosystem, including Copilot and Microsoft 365 Copilot.

[ 🖼️ Image Slot: Скриншот чата и выбора модели (Chat and model selector) ]

The expanded rollout of GPT-5.3 Instant deepens the ongoing integration of OpenAI’s models into the Microsoft ecosystem. Crucially, it marks a pivot away from raw test metrics toward practical, real-world response quality. For enterprise clients, this means faster adoption of AI tools for customer support, IT operations, and internal processes—all without compromising security or compliance.

[ 🖼️ Image Slot: Картинка с призывом подписаться (Subscribe to Newsletter) ]