OpenAI has unveiled its highly anticipated family of language models collectively named GPT-5. The company claims this represents its fastest, smartest, and most efficient model to date. GPT-5 combines a standard intelligence model for most queries with a specialized "GPT-5 thinking" version for complex tasks. The new model is now available to all ChatGPT users.

Concurrently, Microsoft began integrating GPT-5 into its flagship products including Copilot, Microsoft 365 Copilot, the Azure AI Foundry platform, and GitHub Copilot.

In the standard Copilot, users can activate a new "smart mode" where the AI assistant automatically switches between model versions based on task requirements. Like with ChatGPT, GPT-5 capabilities in Copilot are available to all users at no additional cost.

Enterprise customers using Microsoft 365 Copilot have also gained access to GPT-5. Microsoft highlights significant improvements in handling complex tasks, maintaining context in extended conversations, and better understanding user intent. The GPT-5 model will also debut today in the Copilot Studio platform.

Developers receive notable updates as well. GPT-5 has been added to all paid GitHub Copilot tiers, allowing immediate evaluation of its code generation capabilities. The model comes in four variants (gpt-5, gpt-5-mini, gpt-5-nano, and gpt-5-chat), each offering distinct logical processing and code quality characteristics. While GPT-5 excels at multi-step logical problems, GPT-5-chat is optimized for enterprise applications requiring contextual multimodal dialogues.

Finally, Microsoft has incorporated GPT-5 into Azure AI Foundry. This enables developers to leverage OpenAI's flagship model in their AI-powered applications. The platform also introduces a new "model router" feature, ensuring optimal model version selection for specific tasks or queries.

Source