Rumors: OpenAI completes 10T-parameter pretraining of model 'Bel'

Starting August 26, multiple Reddit leakers (Leo's claim relayed by Outside-Iron-8242, plus ChrisGPT, DrSingularity, scaling01, VraserX, TheDigitalHearts, and reposts by ivanbezdomny, haider1, rohanpaulai, imjustnewatai, koltregaskes) posted highly overlapping claims: OpenAI has completed a major pretraining run codenamed Bel with over 10T total parameters, potentially the foundation of its next products. OpenAI has made no response; all information is third-party rumor, and repetition does not equal verification.

Unconfirmed

• Existence and codename: ChrisGPT, self-described as the first leaker, says OpenAI's next major pretraining is codenamed Bel, a 'killer' monster model aimed at outdoing Google's Gemini line; per Leo via Outside-Iron-8242, pretraining is complete.
• Scale: Leakers uniformly cite >10T parameters; scaling01 adds it is roughly the size of GPT-4.5 and that OpenAI internally believes it may even reach the AGI threshold.
• Lineage: scaling01 and DrSingularity call Bel the successor to 'Doug', serving as the base for Astra and GPT-6 (after further RL). Per koltregaskes' relayed sources, after Astra—expected in September, possibly delayed by regulation—Bel is slated for 2026 to surpass 'Fable', described as a 'giant squid'.
• Timing: ChrisGPT expects Bel by year-end or a few months after Astra; koltregaskes' sources point to 2026—accounts conflict.
• Training mechanism: imjustnewatai speculates Bel will never be 'done' training, using fast/slow dual-speed continual learning—fast layers absorbing verified proofs, code and tool traces, slow cycles iterating separately; this is the blogger's personal guess.

Why it matters

• If true, Bel would be OpenAI's key base model for the GPT-6/Astra era, directly shaping its race with Google's Gemini line; rohanpaulai and ivanbezdomny note that compute-constrained Anthropic could stay passive all year.
• Per 新智元, Anthropic plans to release Opus 5.1 this week, boosting per-token intelligence, coding, and long-running agent capability—possibly a response; WorldofAI's roundup also cites Anthropic quietly testing a Claude codenamed Marshmallow and GLM-5.3-Flash appearing.
• VraserX reads this as 'the start of recursive self-improvement' and the singularity—personal speculation.
• Crucially, the leakers' accounts overlap heavily but no primary evidence exists; readers should stay cautious.

2026-08-26 ~ 2026-08-27 · 13 related posts

Full story(2 episodes)→

• Episode 1: OpenAI's next model Astra reportedly solves hard problems, release delayed by risk(2026-08-24, 3 posts)
• Episode 2: Rumors: OpenAI completes 10T-parameter pretraining of model 'Bel'(2026-08-26, 13 posts)

• [source] OpenAI finishes >10T param pretrain 'Bel', claims AGI potential — scaling01 · 2026-08-26
• [source] Rumor: OpenAI Just Finished Next >10T Pretrain "Bel" — Outside-Iron-8242 · 2026-08-26
• Report: OpenAI finished 'Bel', a >10T-parameter pretrain that could underpin post-GPT-6 systems — Dr_Singularity · 2026-08-26
• [source] Leaker Claims OpenAI's Next Big Pretraining Run Is Codenamed 'Bel' — ChrisGPT · 2026-08-26
• OpenAI reportedly completed 10T+ parameter model "Bel" for post-GPT-6 — VraserX · 2026-08-26
• Scoop: OpenAI Finishes GPT Bel Pretrain, Theory on Continuous Learning — imjustnewatai · 2026-08-26
• Rumor: Anthropic to Release Opus 5.1 This Week; OpenAI Pre-Trains 10T Parameter Model — 新智元 · 2026-08-26
• GPT-6 'Bel' pretrain reportedly done at 10T+ params; GLM-5.3-Flash unmasked — WorldofAI · 2026-08-27
• Rumor: OpenAI's post-GPT-6 model codenamed 'Bel', targets 2026 release — koltregaskes · 2026-08-27

4 near-duplicate retellings: rohanpaul_ai · ivan_bezdomny · haider1 · The_Digital_Hearts