Report: ByteDance Pre-Training 10-Trillion Parameter AI Model, Far Exceeding Kimi
eyishazyer · x · 2026-08-07
According to a report by the Financial Times (FT), ByteDance is currently pre-training an AI model with up to 10 trillion (10T) parameters, vastly exceeding the scale of the recently popular Kimi K3 (2.8T parameters).
The report also notes that ByteDance has avoided distilling rival models for over a year, preferring independent development. If this reported run succeeds, it will demonstrate ByteDance's ability to execute frontier-scale pre-training without leaning on a competitor's model as a teacher.
Related event: Report: ByteDance is Training a 10-Trillion Parameter AI Model(18 posts)→
More from Models
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16
- Argon edges out astra as daily driver, but Claude still wins heavy coding work — iruletheworldmo · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- RoboVista benchmark: Gemini-4 Argon hits 81.9% on robotics understanding — berkeley_ai · 2026-10-16
- Dev reports decision model jev beats gpt-6-luna on quality in real training pipeline — genmon · 2026-10-16
- Google won't publish Argon 4 limits in AI Pro—users compare experiences — Consistent-Text-2662 · 2026-10-16