Scoop: OpenAI Finishes GPT Bel Pretrain, Theory on Continuous Learning
imjustnewatai · x · 2026-08-26
A scoop claims OpenAI has finished its next pretrain, codenamed "Bel", with over 10T parameters, potentially serving as the base for GPT-6 or an AGI-threshold model. The author theorizes that Bel will utilize a two-speed learning mechanism: a fast layer absorbing lessons from verified proofs and tool traces, and a slower loop consolidating successful improvements into persistent weights. This theory is supported by the
Related event: Rumors: OpenAI completes 10T-parameter pretraining of model 'Bel'(13 posts)→
More from Models
- Reasonable post-trains Nemotron 3.5 on 3,000 Verus proofs, lifting pass@1 ~30 points to 73.3% — fhuszar · 2026-10-16
- NVIDIA's Nemotron 3.5 Super VL hits CoreWeave: 12B-active multimodal MoE at $0.25/M input tokens — wandb · 2026-10-16
- Opus 5 might be a very deep model or use looping, commentator speculates — scaling01 · 2026-10-16
- GPT-6.1 Sol audit finds 6% of arXiv papers had false main theorems, down to 1.5% recently — nitarshan · 2026-10-16
- Nvidia-backed post-training pushes Nemotron 3.5 VL past frontier on dermatology diagnosis — LoganGrasby · 2026-10-16
- Gemini 4 Argon scores 97.5% on ARC-AGI-3, second lab above 95% — patience_cave · 2026-10-16