Tencent says Hy4 flash processed over 50 trillion tokens on its first day

As reported on November 18, 2026, Tencent said its Hy4 flash language model processed more than 50 trillion tokens on its first day of availability—a first-day record for its Hunyuan model series. Tencent also said Hy4 flash topped the daily large-language-model ranking on OpenRouter that day.

Hy4 flash was released on November 16. Tencent said first-day usage on its WorkBuddy platform was 10 times that of Hy4 preview’s first day, with the model completing more than 50 million tasks. The company reported a generation speed of about 200 tokens per second and peak processing of more than 1 billion tokens per second.

According to Tencent, Hy4 flash has 224 billion total parameters, including 64 billion Engram parameters, with 12 billion active parameters. It supports multimodal input and a 1 million-token context length. Tencent lists pricing at 0.5 yuan per million input tokens, 2 yuan per million output tokens and 0.02 yuan per million tokens when a cache is hit—substantially lower than Hy4 pricing.

Tencent’s Hunyuan team said Hy4 flash performed slightly better overall than the larger Hy4 preview while costing less. On Tencent’s internal WorkIndex evaluation, which covers tasks including presentations, documents, spreadsheets, audio and video, and visual understanding, the team put its per-task cost at one-thirteenth of Hy4 preview’s and one-twentieth of Hy4’s. Those performance and cost comparisons are Tencent’s own assessments.

Further reading