Leaked GitHub Documents Reveal GPT-5 AI Models, Features, and Availability Ahead of Official Launch

A recently deleted GitHub blog post revealed details about four variants of OpenAI’s upcoming GPT-5 AI model.

OpenAI’s upcoming AI model, GPT-5, has reportedly been leaked ahead of its official release through a now-deleted GitHub blog post. The post revealed details about four variants of GPT-5, including the base GPT-5 model, GPT-5 Mini, GPT-5 Nano, and a new GPT-5 Chat tailored for enterprise use. Additionally, a GPT-5 Pro model is expected to be available exclusively for OpenAI’s Pro subscribers. Alongside these models, ChatGPT is said to receive a redesigned interface.

The leak was first spotted by TestingCatalog before the post was taken down, but screenshots quickly circulated on social media. The Mini variant is believed to offer lower latency, while the Nano version might be optimized for edge or local deployments. Though details about GPT-5 Thinking are sparse, it is speculated to be an API that lets developers adjust the model’s “thinking budget.”

GPT-5 Chat is designed for integration with tools like GitHub Copilot, promising more advanced, natural, and context-aware conversations tailored for enterprise applications. This variant aims to support multimodal inputs and deliver deeper conversational understanding.

Interestingly, the leak suggests that GPT-5 will be made available even to users on the free tier, marking a significant expansion in access to OpenAI’s most advanced language model.