ByteDance goes big – very big – on AI
This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).
ByteDance is training an AI model with up to 10 trillion parameters, which would make it comparable in scale to Anthropic's Mythos system (estimated at ~8 trillion parameters). This dwarfs other Chinese competitors: Moonshot AI's Kimi K3 has 2.8 trillion parameters, while Meituan's LongCat-2.0 and DeepSeek's V4-Pro previously led China with 1.6 trillion. The model is currently in pre-training, a process that typically takes three to six months before fine-tuning and release. Direct comparisons with US models remain difficult as Anthropic and OpenAI do not publicly disclose parameter counts.
Table of contents
Read: China weighs blocking the West from its best AIQuestions this post answers
How many parameters does ByteDance's new AI model have and how does it compare to other models?
ByteDance is training a model with up to 10 trillion parameters. That is more than three times the size of Moonshot AI's Kimi K3 (2.8 trillion parameters), which previously led China. Industry estimates put Anthropic's Mythos 5 at around 8 trillion parameters and Fable 5 at about 5 trillion, making ByteDance's model comparable to Mythos in scale. Developers tracking the AI model arms race between Chinese and US labs follow developments like these on daily.dev.