TellWorldAI World Archive
Ask中文

Home›Models

Model

Hy4 flash

Hy4 flash is Tencent's most cost-effective Hy model to date, with 224 billion total parameters, 12 billion active, multimodal input and up to one million tokens of context.

Status AvailableModality Text, ImageContext (K tokens) 1000 Era Generative AI for everyone

Parameters
224B
Access
API, App
Release date
Developer
Tencent

What it is

Hy4 flash is a model from Tencent, which describes it as the most cost-effective Hy model to date.[1] It has 224 billion total parameters, including 64 billion Engram parameters, and activates 12 billion parameters.[1] It supports multimodal inputs and a context length of up to 1M tokens.[1]

Origins

Tencent announced Hy4 flash on November 16, 2026, and positions it for everyday workflows.[1] Tencent says the model slightly outperforms the larger Hy4 Preview, which has 770 billion total parameters and 49 billion active, on Tencent's internal WorkIndex benchmark.[1]

Main features

Tencent says the KV cache is 1/24 the size of Hy3's and 1/6 the size of Hy4's.[1] The KV cache stores intermediate attention data during generation, so a smaller cache lowers the memory a request needs. Tencent also says reinforcement learning cut average end-to-end task output tokens by 50%, and that decoding exceeds 150 tokens per second.[1]

On the WorkIndex benchmark, Tencent says the per-task cost of Hy4 flash is 1/13 that of Hy4 Preview, 1/20 that of Hy4, and 1/8 that of DeepSeek-V4.1-Flash-1017 at its Max tier under peak-hour pricing.[1] These are Tencent's internal figures.

Tencent's approximate API prices are US$0.070 per million input tokens, US$0.281 per million output tokens and US$0.003 per million cached input tokens.[1]

Position in development

Hy4 flash is available through Tencent Cloud TokenHub, WorkBuddy, Yuanbao, WeCom, QQ, ima and Marvis.[1] Tencent says it plans to open-source Hy4 flash, with frequent iterations planned, and has not given a release date.[1]

CapabilitiesLarge language modelsMultimodal

ApplicationsEnterprise

Ecosystem & regionChina

20261

  1. ReleaseTencent launches Hy4 flash, a cheaper 1M-context multimodal model

    Tencent announced Hy4 flash on 16 November 2026 as its most cost-effective Hy model to date, with up to one million tokens of context and API prices of about US$0.070 per million input tokens.

Developed by1

Tencent

Notes and sources

  1. Tencent launches Hy4 flash to power high-speed, cost-efficient AI deployment for everyday workflowshttps://www.tencent.com/tencent-launches-hy4-flash-to-power-high-speed-cost-efficient-ai-deployment-for-everyday-workflows/

Cite this page: Hy4 flash. TellWorld, https://www.tellworld.net/models/hy4-flash. Updated 2026-11-18