OpenAI set to launch GPT-5 in August with new mini and nano variants: report
text_fields
OpenAI is reportedly gearing up to launch its next-generation artificial intelligence model, GPT-5, in August.
According to a report by The Verge, the upcoming model will introduce three variants - standard, mini, and nano - marking a significant shift in the company’s strategy.
The release is expected to mark the end of the o-series models, with GPT-5 aiming to unify the GPT and o-series into a single, advanced model. This unified model is expected to feature native reasoning capabilities.
Speculation about GPT-5 has been circulating since last year, initially under the codename “Orion.”
However, that codename was later confirmed to be associated with GPT-4.5, which launched in February. In the same month, OpenAI CEO Sam Altman revealed a detailed roadmap for the future of its models on X (formerly Twitter).
Altman noted that GPT-4.5 would be the final model without chain-of-thought (CoT) capabilities. “All future GPT-series models will natively know when to think for a long time or not,” he stated, emphasising a shift toward models that can determine the complexity of reasoning required based on context.
Altman also shared that GPT-5 would be available to all ChatGPT users at a standard intelligence level, while Plus and Pro subscribers would gain access to more advanced capabilities.
The new model will also integrate with OpenAI’s full suite of tools, including voice, Canvas, Search, and Deep Research.
Following months of silence, OpenAI’s Member of Technical Staff, Alexander Wei, recently posted, “we are releasing GPT-5 soon,” sparking renewed anticipation.
According to The Verge, the GPT-5 release will follow the debut of OpenAI’s upcoming open-source models. The base and mini variants of GPT-5 will be accessible through both ChatGPT and the API, while the nano variant will be exclusive to API users.
