Skip to content
Hy4 flash Hub

Comparison · checked 17 Nov 2026

Hy4 flash vs DeepSeek V4.1 Flash

A sourced side-by-side of Tencent’s Hy4 flash and DeepSeek’s current Flash API (model name deepseek-flash, version DeepSeek-V4.1-Flash-1017). Prices are list rates unless a dated promotion still applies.

The short version

A

Hy4 flash

by Tencent

A 12B-active Hy4 variant with a flat TokenHub list price of $0.070 / $0.281 per 1M tokens.

Choose it for

  • Everyday office work, per Tencent’s own positioning
  • A single list rate, with a dated 50% TokenHub promo through 30 Nov 2026
  • Calling through TokenHub, OpenRouter or Vercel AI Gateway today
Tencent Hy4 flash post

B

DeepSeek V4.1 Flash

by DeepSeek

The DeepSeek API’s Flash model (deepseek-flash / V4.1-Flash-1017): 1M context, 384K max output, peak and off-peak prices.

Choose it for

  • Direct DeepSeek API use with a 384K max output cap
  • Workloads that can wait for off-peak hours
  • Vision, tools and thinking mode on DeepSeek’s own docs
DeepSeek pricing

VerdictUse Hy4 flash when you want Tencent’s flat TokenHub rate or you already work in Tencent products. Use DeepSeek-V4.1-Flash when you want DeepSeek’s API, the 384K output cap, or off-peak discounts. Tencent’s claim that Hy4 flash costs less per WorkIndex task than DeepSeek-V4.1-Flash-1017 at Max-tier peak-hour pricing is a vendor benchmark, not an independent leaderboard.

Quick answer

Both are hosted Flash-class APIs; the bill and the output cap are the clear differences

DeepSeek’s pricing page, checked 17 November 2026, says the current Flash API model name is deepseek-flash and the version is DeepSeek-V4.1-Flash-1017. It lists a 1M context window, a 384K maximum output, vision, tool calls, and thinking mode (on by default, with a documented way to switch). Peak hours are 01:00–04:00 and 06:00–10:00 UTC, Monday through Friday, excluding Chinese public holidays; off-peak rates are half of peak.

Hy4 flash, on Tencent TokenHub, also has a 1M context window, but the API guide caps input at 960k tokens and output at 64k tokens. List prices do not split peak and off-peak. Tencent’s launch post compared average WorkIndex task cost with DeepSeek-V4.1-Flash-1017 (Max tier, peak-hour pricing) and said Hy4 flash was one-eighth of that cost. We quote that sentence as a vendor claim only.

Matrix

Published limits and list prices

ItemHy4 flashDeepSeek-V4.1-Flash-1017
API model idhy4-flashdeepseek-flash
Context window1M tokens1M tokens
Max output64k (TokenHub API guide)384K (DeepSeek pricing page)
Input price / 1M, cache miss$0.070 list$0.15 off-peak / $0.30 peak
Output price / 1M$0.281 list$0.60 off-peak / $1.20 peak
Cache-hit input / 1M$0.003$0.003 off-peak / $0.006 peak
Open weights (17 Nov 2026)Promised, not publishedSee DeepSeek’s own model releases; not claimed here

Decide

Choose by the job and the bill

Lean toward Hy4 flash if

  • You already have a Tencent Cloud TokenHub key, or you work in WorkBuddy / Yuanbao
  • You want a flat list rate rather than peak and off-peak windows
  • Your outputs fit inside a 64k-token cap

Lean toward DeepSeek V4.1 Flash if

  • You need the 384K maximum output documented by DeepSeek
  • You can schedule jobs in DeepSeek off-peak hours
  • You want to bill against DeepSeek’s own API and pricing page

Check the live prices

Open Tencent’s TokenHub pricing page and DeepSeek’s pricing page, then run the token estimator if you need a rough monthly figure.