The company says a new internal model has produced results on more than 100 open problems, prompting outside mathematicians to advise on review and publication.
The company is increasing spending on chips, models and computing capacity as it pushes deeper into artificial intelligence.
The proposed agreement would let the rivals probe each other’s commercial systems for safety failures as scrutiny of frontier AI grows.
Paid subscribers accuse Anthropic, OpenAI, Google and SpaceXAI of illegally coordinating efforts to slow AI development.
The US president says his administration will not slow AI development as leading technology executives call for more time to strengthen safeguards.
