ByteDance is reportedly in the early stages of training an AI model with up to 10 trillion parameters, a scale that could rival Anthropic's most advanced Mythos system. If completed, it would be roughly three times larger than Moonshot's Kimi K3, currently the largest Chinese model released.
The pre-training phase typically lasts three to six months, followed by fine-tuning and a potential release, though the final model size is not yet fixed. The effort highlights how Chinese tech firms are rapidly closing the gap with leading U.S.
AI labs.
Comments