Vibeleaderboard
← All Intel
Intel / article

Hy4 Flash now available on AI Gateway for 50% off

Source
vercel.com
Author
Zachary Chen
Date

Context

Hy4 Flash, the low-cost model in Tencent's Hy4 family, is now on Vercel's AI Gateway at half price on input, cached input and output tokens through November 29, 2026. AI Gateway is Vercel's single API for calling models from many companies, and it passes through each provider's price with no markup. Tencent's own launch post lists Hy4 Flash at $0.070 per million input tokens, $0.281 per million output tokens and $0.003 per million cached tokens. Vercel's model page currently shows $0.035 and $0.1405 for input and output, exactly half the list price, so for the next two weeks the discount is the whole difference between buying from the gateway and the list rate.

Vercel describes Hy4 Flash as keeping the full Hy4 feature set at roughly a tenth of the price per token. The numbers bear that out: Hy4 lists at $0.834 per million input tokens and $2.501 per million output, so Flash's list price is about 8% of Hy4's for input and 11% for output. The shared features include reasoning, which the model can keep across the turns of a conversation instead of starting fresh each time, along with , structured outputs that follow a set format, , image input and the same 1M-token . Reasoning can also be switched off for requests that do not need it, which Vercel pitches at high-volume coding, and document work, where a model that answers directly spends fewer tokens.

What the discount does not settle is how good the model is. Tencent claims Hy4 Flash matches Claude Haiku 5.5 at high reasoning effort on everyday work in human blind tests, with an 18-fold reduction in task cost, but that is the company's own evaluation and no independent benchmark of Flash has appeared in Intel yet. The full Hy4, by contrast, has been measured independently by Artificial Analysis, the source of the list price above. Hy4 itself arrived on AI Gateway earlier this month with its own 50% discount running through December 31. For anyone already routing work through Vercel, Flash is a cheap model to try during the discount window; whether it can replace a Haiku-class model is something to test on your own tasks rather than take from the launch claim.

Terms in this piece · Glossary
  • tool use — A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
  • prompt caching — Reusing the model's processed form of a repeated prompt prefix so subsequent calls skip re-reading it, cutting cost and latency substantially.
  • context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
  • AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
More from Zachary Chen
Recommended reads
Comments

Checking sign-in…

Loading comments…