ByteDance Weighs 5-Trillion-Parameter AI Model to Challenge Rivals

ByteDance has stepped up investment in generative artificial intelligence, with its Seed Foundation team emerging as a key hub for foundation-model research. A model exceeding 5 trillion parameters would be substantially larger than systems associated with Alibaba’s Qwen and Moonshot AI’s Kimi. The proposed scale underscores how Chinese technology companies are using computing capacity, model size and development speed to compete for influence in domestic and global AI markets.

As of Aug. 9, 2026, ByteDance was discussing plans for Seed Foundation to lead development of a large language model with more than 5 trillion parameters, according to reports. The targeted size would put it near Anthropic’s reported Mythos model and ahead of prominent Chinese rivals by parameter count. The proposal remains under discussion, with ByteDance yet to disclose a training schedule, computing requirements or the amount it plans to invest.