OpenAI has released GPT-6 Astra to a limited set of organizations, with wider access through paid ChatGPT plans and cloud platforms coming within days.
Physical Superintelligence launched with $58 million in seed funding to build an AI physics lab and develop Emmy, a platform for virtual physicists. Its first commercial use targets data center optimization.
Skywork AI detailed Matrix-Game 3.5, an interactive world model with Patch Memory, 720P streaming video generation at about 20 FPS on a single GPU, and one minute memory.
Safe Superintelligence Inc. has entered a strategic partnership with NVIDIA that will expand its compute capacity by 10x. NVIDIA also made an investment in the AI lab, which is led by Ilya Sutskever.
Ineffable Intelligence has partnered with Google Cloud to develop a continuously learning AI system called the 'superlearner'. The collaboration will use Google Cloud's AI infrastructure, including NVIDIA Vera Rubin NVL72 GPUs, to support large-scale reinforcement learning research.
Abu Dhabi-based G42 and the Government of India have finalized the commercial framework for Condor Galaxy India, an 8-exaflop AI supercomputer comprising 64 Cerebras CS-3 systems. The agreement was signed during Prime Minister Narendra Modi's state visit to Abu Dhabi.
PATEO CONNECT Technology and Aumovio Holding China have signed a memorandum of understanding to jointly develop high computing platforms and artificial general intelligence technologies for software defined vehicles, focusing on global market expansion.
OpenAI has disbanded its mission alignment team, with members reassigned across the company, while former team head Josh Achiam becomes chief futurist.
OpenAI has introduced GDPval, a new benchmark to assess AI models like GPT-5 against human professionals across various industries. The evaluation spans 44 occupations and aims to measure AI's performance on economically valuable tasks.
Meta has paused hiring in its AI division following a significant recruitment drive, according to Reuters. This move comes as part of a broader restructuring of its 'Superintelligence Lab.'
Litmus7 Inc has launched an AGI innovation hub in Trivandrum, India, as part of its 'Advait' initiative, focusing on developing human-centered AI called 'Neural Intelligence'.
DeepMind has announced Genie 3, a new world model designed to train AI agents in interactive environments, marking a step towards artificial general intelligence.
Meta CEO Mark Zuckerberg indicates a shift in the company's approach to open sourcing its AI models, focusing on safety concerns and monetization strategies.
Microsoft is in advanced talks with OpenAI to secure ongoing access to its technology, even if OpenAI achieves artificial general intelligence (AGI), according to Bloomberg. The negotiations aim to redefine their partnership as OpenAI transitions to a for-profit model.
OpenAI is preparing to release its next-generation AI model, GPT-5, potentially as early as August. The new model will integrate distinct models, including the o3 model, to perform a wider range of functions.
OpenAI has announced the impending release of GPT-5, while clarifying that a separate experimental model was used at the IMO. The new model incorporates advanced research techniques.
The Kenya Private Sector Alliance (KEPSA) and MindHYVE.ai have signed a strategic agreement to deploy AGI agents for workforce and SME development in Kenya.
OpenAI has decided to keep its nonprofit in control, reversing plans to become a for-profit entity. The decision follows discussions with state attorneys general and addresses concerns about maintaining its mission.
Anthropic has initiated a research program to investigate the welfare of AI models, considering their potential consciousness and experiences.
Safe Superintelligence, co-founded by Ilya Sutskever, has secured $2 billion in funding, boosting its valuation to $32 billion, despite not yet having a product.
Daily report covering major AI developments and industry news, with both top stories and complete market updates
