OpenAI’s new GPT-5 models announced early by GitHub

GitHub has accidentally announced OpenAI’s new range of GPT-5 models. A now-deleted GitHub blog post reveals that GPT-5, which will be available in four different versions, offers “major improvements in reasoning, code quality, and user experience.”

Reddit posters noticed a new GitHub blog post announcing that GPT-5 is generally available in GitHub models, before the blog post was quickly removed. An archived version reveals that GPT-5 will have “enhanced agentic capabilities” and can handle “complex coding tasks with minimal prompting.”

GitHub also compared GPT-5 to Llama 4 Scout and Cohere v2.

Image: GitHub

GPT-5 will have four model variants, according to GitHub:

gpt-5: Designed for logic and multi-step tasks.
gpt-5-mini: A lightweight version for cost-sensitive applications.
gpt-5-nano: Optimized for speed and ideal for applications requiring low latency.
gpt-5-chat: Designed for advanced, natural, multimodal, and context-aware conversations for enterprise applications.

I revealed last month that OpenAI was preparing to launch GPT-5 in early August, complete with mini and nano versions that will also be available through its API. Earlier this week multiple OpenAI employees started teasing that it would be a big week, and OpenAI CEO Sam Altman even teased GPT-5 over the weekend.

OpenAI has all but confirmed it’s launching GPT-5 later today, with the announcement of a “LIVE5TREAM” for 10AM PT / 1PM ET. OpenAI’s GPT-5 launch comes in the same week the company also delivered two GPT-OSS open-weight models, with one small enough to run locally on a PC.