Frontier models are being credited with results on named open problems in math — a concrete reference point for how far model-assisted research reasoning has moved, and for what to expect on hard verification-heavy work.

• postChatGPT Business gets a $100 Premium seat tier
• postPreviewing Ultrafast mode: GPT-5.6 Sol at up to 14x the speed. Launching first i
• postChatGPT starts keeping a record of your desktop work
• posthttps://t.co/zBh8ck0Gbc

• articleTen advances in mathematics and theoretical computer scienceopenai.com
• articleTen advances in mathematics and theoretical computer sciencesimonwillison.net
• videoEinstein Arena: Harnessing Collective Agent Intelligence for Open Science — James Zou, Together AIAI Engineer

Checking sign-in…

Loading comments…