OpenAI finishes >10T param pretrain 'Bel', claims AGI potential
scaling01 · x · 2026-08-26
- Scoop: OpenAI finished pretraining "Bel", the successor to "Doug", with >10T total parameters, similar in size to GPT-4.5.
- Positioning: Expected to be the base for Astra and GPT-6. OpenAI believes it could be a post-GPT-6 base or even an AGI-threshold model.
- Competition: OpenAI assesses that Anthropic lacks a strong public response to Astra due to compute constraints. Anthropic expects OpenAI to dominate the rest of the year but plans to return to form early next year.
Related event: Rumors: OpenAI completes 10T-parameter pretraining of model 'Bel'(13 posts)→
More from AGI Musings
- UMD Philosophy hires: junior Philosophy of Mind and joint AI Ethics/Safety positions — sethlazar · 2026-10-16
- Dwarkesh: Anthropic insider expects AI to match humans at AI research within 9 months — ZeroStateReflex · 2026-10-16
- Michigan professor Ben Green: why 'AI for Good' fails and self-regulation won't work — benzevgreen · 2026-10-16
- "Everything Is an Alignment Problem": AI Models, Business Models, Teams, Providers — saranormous · 2026-10-16
- David Brin: LLMs lack 'wanting,' evolution's key product, in new machine consciousness book — cccalum · 2026-10-16
- AI doesn't kill the joy of learning, it kills the prestige of it — ZeroStateReflex · 2026-10-16