OpenAI will officially release the highly anticipated GPT-5 AI model series today at 10 a.m. Pacific Time (August 7). However, key details of the model were accidentally leaked in advance through the Microsoft GitHub repository.
A now-deleted GitHub blog post confirmed that GPT-5 will be available on GitHub Models. While there had been speculation that OpenAI might only preview GPT-5 and not officially launch it today, this blog post debunks those rumors and confirms that the GPT-5 model will indeed be released today.
The GitHub post states that GPT-5 will be OpenAI’s most advanced model to date, bringing significant improvements in inference, code quality, and overall user experience. It will be able to handle complex coding tasks with minimal input, provide more accurate code interpretation, and demonstrate improved agent capabilities.
GPT-5 will be available in the following four versions:
GPT-5: Designed for logical and multi-step tasks.
gpt-5-mini: lightweight version for cost-sensitive applications.
gpt-5-nano: Optimized for speed, ideal for applications requiring low latency.
GPT-5-chat: Designed to deliver advanced, natural, multi-modal and context-aware conversations for enterprise applications.
Developers will be able to access GPT-5 in the Models tab and explore its capabilities in the GitHub Model Playground or through the GitHub API.
For regular consumers, OpenAI’s GPT-5 will be available through ChatGPT and Microsoft’s Copilot. In Copilot, GPT-5 will run in a new "smart" mode. In this mode, Copilot can adjust the speed of its thinking, whether it's faster or deeper, depending on the task at hand. While seemingly simple on the surface, this feature is unique to the new GPT-5 model.
On ChatGPT, free users will have unlimited access to GPT-5 models at the standard intelligence level. Plus level users can choose to run GPT-5 at higher intelligence settings, while Pro level users can use GPT-5 at more advanced intelligence levels.