Five major Chinese AI labs slashed token prices 50–99%, signaling that model capabilities have converged and price is now the only competitive lever.

What happened: ByteDance cut Seedance 2.0 Mini to 23 yuan ($3.40) per million tokens; Tencent slashed Hy-MT2-Pro by nearly 70%; MiniMax halved its M3 series; Alibaba cut Qwen3.7-Max 50% for the 618 shopping event; and Xiaomi dropped MiMo V2.5 by 99%.

Why it matters: Bank of America Securities analysts attribute the pricing race to limited capability gaps across incumbents, meaning Chinese AI model APIs are moving from differentiated products to a commodity layer where only cost drives choice. For enterprises and developers using these APIs, the 50–99% price drops sharply lower inference costs, but for smaller labs without major cash reserves like ByteDance or Alibaba, the sustained margin pressure poses an existential risk.

What to watch: Whether ByteDance's 23 yuan ($3.40) per million token price for Seedance 2.0 Mini is sustainable, and whether Alibaba's Qwen3.7-Max discount tied to the 618 event signals a permanent price floor or a temporary promotion.