Leak reveals GPT-5's features and variants hours before official launch
What's the story
GitHub has inadvertently announced the latest range of OpenAI's GPT-5 models. The now-deleted blog post highlighted that the upcoming AI model will be available in four different variants and promises "major improvements in reasoning, code quality, and user experience." The post was quickly removed but not before Reddit users managed to capture an archived version detailing the capabilities of these new models.
Enhanced capabilities
Major improvements in coding tasks
The archived blog post from GitHub also revealed that the new GPT-5 models will have "enhanced agentic capabilities" and be able to tackle "complex coding tasks with minimal prompting." This indicates a significant leap in the AI's ability to understand and execute intricate programming assignments, making it an even more valuable tool for developers and tech enthusiasts alike.
Model variants
Four different models under the new range
The GitHub blog post detailed the four different models that will be part of the new GPT-5 range. They are: 1. gpt-5: Aimed at logic and multi-step tasks. 2. gpt-5-mini: A lighter version for cost-sensitive applications. 3. gpt-5-nano: Speed-focused and ideal for low-latency applications. 4. gpt-5-chat: For advanced, natural, multimodal, and context-aware conversations in enterprise applications.
Anticipated release
GPT-5 launch scheduled for today
Just a month ago, news broke that OpenAI was gearing up to unveil GPT-5 in early August. This would include mini and nano versions available via its API. The company has all but confirmed the launch for later today with a "LIVE5TREAM" scheduled for 10:30pm IST. The announcement comes just days after OpenAI introduced two GPT-OSS open-weight models, one of which is small enough to run locally on a laptop with 16GB RAM.