August 8, 2025 — OpenAI has released GPT-5, calling it its most capable AI system to date, and is rolling it out to all ChatGPT users. The new model introduces a unified approach that can answer quickly for simple prompts or “think longer” on harder tasks, aiming to deliver expert-level responses without users having to pick a separate mode.
In parallel, Microsoft said GPT-5 is available today across its Copilot lineup — including Microsoft 365 Copilot and Copilot Studio — with a behind-the-scenes router that chooses the best model for each task. The company describes improved context handling and sustained conversation quality as part of the upgrade.
For developers, OpenAI is releasing GPT-5 via API in three sizes — gpt-5, gpt-5-mini, and gpt-5-nano — alongside new controls for reasoning effort and verbosity. The developer post also details big gains on coding and agentic benchmarks, and clarifies that the ChatGPT experience uses routing between fast and deep-reasoning components.
GitHub confirmed a public preview of GPT-5 in GitHub Copilot for paid subscribers, highlighting better reasoning, code quality and end-to-end task handling with clearer explanations of what the model is doing.
OpenAI positions GPT-5 as a significant step forward in general capability, with state-of-the-art results across coding, math, writing, health and visual perception — and a chat experience that dynamically shifts between speed and deeper reasoning as needed. The rollout to free and paid ChatGPT tiers is underway.