AI reasoning model tops key benchmarks, outperforming proprietary reasoning models from rivals like Google and OpenAI.

Although there is currently no public timeline for the model's release, the development signals a possible recalibration of OpenAI's position in an increasingly crowded and fast-moving space.

Thinking-2507, as we'll call it for short, now leads or closely trails top-performing models across several major benchmarks.

Chinese AI firm Zhipu, rebranded as Z.ai, reportedly plans to release its GLM-4.5 model next week, enhancing its competitive ...

OpenAI says it needs more time to conduct safety checks before releasing its first open-weight model. There’s no confirmed timeline for launch yet.

Microsoft and OpenAI are in advanced talks to rewrite their partnership, resolving the 'AGI clause' to ensure continued ...

OpenAI is currently working on new office tools, which are to be added to its AI model, ChatGPT, according to a new report.

Top members of Meta’s new Superintelligence Lab discussed pivoting away from the company’s powerful open source AI model, ...

Meta investors will be scrutinizing CEO Mark Zuckerberg’s AI hiring blitz during the company’s second-quarter earnings on ...

Meta is weighing a dramatic shift in its AI strategy, as its newly formed Superintelligence Labs considers abandoning its most powerful open-source AI model in favour of developing a closed ...