AI TrackerGet the app
← All news
Models

Tencent releases Hy4 Flash, priced at $0.07 per million input tokens

Tencent Hunyuan announced Hy4 Flash, a smaller sibling of Hy4, which it calls the cheapest model for everyday productivity. Tencent says that in human blind tests on daily work it matches Claude Haiku 5.5 (high) at an 18x lower task cost; these are Tencent's claims, and the test details are on its research page at hy.tencent.ai/research/hy4-flash.

Tencent's API pricing, per million tokens:

  • Input: $0.070
  • Cached input: $0.003
  • Output: $0.281

A reply from OpenCode says the model is free there for two weeks, with a 1M-token context window. On November 16, OpenRouter's catalog also listed the model as tencent/hy4-flash, with a context length of 1,048,576 tokens.

Get news like this as a push.

AI Tracker alerts you when something changes for the tools you use.

Get the app