OpenAI Strawberry, a rumoured new artificial intelligence (AI) model with enhanced reasoning and mathematical problem-solving capabilities, will reportedly be released by the end of September. As per a report, the new AI model will be launched within ChatGPT as well as be included in a separate offering. Earlier this year, the AI firm was rumoured to be working on this new generative AI technology which can think and solve complex tasks that require multi-step problem solving. Notably, Strawberry is said not to be the next generation of GPT but a separate model with added features.
OpenAI Nearing the Release of Strawberry AI
The Information reports that the purported AI model is currently in the final phase of internal testing, which usually suggests a beta or public release is on the way. Citing unnamed people who tested the model, the publication claimed the Strawberry model could arrive in the next two weeks.
However, it also added that plans could change since release timelines are usually subject to change. Notably, Strawberry is the internal codename of the AI model, and it could be introduced with a different branding at launch.
Additionally, Strawberry AI is said to have “the ability to think”. The AI model reportedly does not immediately start generating a response once the prompt has been added, and takes between 10 to 20 seconds before providing the output.
The AI model is said to be available within the company’s native chatbot ChatGPT. However, it will reportedly be available as a separate offering, likely for those clients who use the application programming interface (API) version of the tool. Details of how these two iterations will be integrated and whether they will have any differences are not known.
At launch, the AI model will only be text-based, with no support for images or videos, the report claimed.
Last month, a report revealed that OpenAI’s Strawberry AI could be released in the fall, but did not specify any particular timeline. The report also highlighted that the AI model will not be presented as GPT-5, as the next iteration of the large language model is under development.