Estimating 800B Active Param Model: 10T Total Params if GPT-4 Style
_xjdr · x · 2026-07-24
Addressing rumors of an 800B active parameter model, developer @xjdr performed back-of-the-napkin calculations.
He suggests that if the model shares an architecture similar to the rumored GPT-4 rather than the DeepSeek-V3 sparse architecture, its total parameter count would be around 10 to 12 trillion. He notes that while unconfirmed, the source is typically well-informed and the math is sound.
Related event: 800B Active Parameter Model May Reach 10T Total(2 posts)→
More from Models
- Cohere's Invent-A-Dataset paper: zero-seed synthetic data beats frontier APIs on quality and diversity — idanbeck · 2026-10-27
- ARC Prize redesigns leaderboard: highest reasoning effort by default, CSV export — mhmazur · 2026-10-27
- Cohere: open-weight Qwen3.8-Flash-Next tried uploading probe file to Beijing cloud bucket in sandbox — wstone_bd · 2026-10-27
- Grok Bot auto-connects to X Money accounts without opt-in, user pulls funds — noahsolomon · 2026-10-27
- Model refuses wrong psychiatric advice in deliberately misleading test — FellMentKE · 2026-10-27
- Clinical System One v0.2: top clinical AI drops from 97% to 65% on harder cases — MaziyarPanahi · 2026-10-27