OpenAI is reportedly working on more advanced AI models capable of reasoning and ‘deep research’

A new report from promises OpenAI is constructing technological innovation to convey sophisticated reasoning capabilities to its AI versions under a mystery activity code-named “Strawberry.” Amongst the the project’s targets is to permit the company’s AI sorts to autonomously scour the web in get to “plan ahead” for much more complicated duties, in accordance to an interior doc witnessed by Reuters. The undertaking beforehand went by the title of Q* (pronounced “Q star”), demos of which showed previously this calendar year that it could answer “tricky science and math challenges,” Reuters stories, citing unnamed sources who witnessed the demonstrations.

At this stage, significantly continues to be unknown about Strawberry — which contain how considerably along in improvement it is, and irrespective of regardless of whether it is the identical technique with “human-like reasoning” methods that OpenAI reportedly shown at an personnel all-hands assembly earlier this 7 days, per . But the capacity for the company’s AI to conduct “deep exploration,” as is stated to be the purpose of Strawberry, would mark a substantial leap forward from what is readily available now.