---
format: "aidr-story-markdown/v1"
id: "d369c4d2e8066d06ff5dc375669ce33fdd6c6bdcb2a16ed0c5186e83d0ebd462"
canonical_url: "https://aidr.today/d369c4d2?lang=en"
title: "Tencent Releases Open Weights for Hy4 Flash Multimodal AI Model"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-18T16:52:00.000Z"
category: "Models"
topics: ["open-source"]
source_urls: ["https://marketbrief.now/ai/tencent-releases-open-weights-for-hy4-flash-multimodal-ai-model-e35843f3","https://x.com/TencentHunyuan/status/2122912575302119554","https://x.com/TencentHunyuan/status/2122217770314301469","https://x.com/vllm_project/status/2122914790377164886","https://x.com/vllm_project/status/2122917809948852559","https://x.com/Sino_Market/status/2122979364631720140","https://x.com/himanshustwts/status/2122319244042047969","https://x.com/martinvars/status/2122399510882074750"]
summary: "Tencent released the open weights for its Hy4 Flash multimodal AI model, including FP8 and Base versions. The model uses a mixture-of-experts architecture with 224B total parameters and 12B active parameters per token, supporting a 1M token context window. The weights are distributed under the Tencent Hy License. The release coincides with day-0 support from the vLLM and SGLang runtimes. Tencent claims Hy4 Flash slightly outperforms its larger 770B parameter Hy4 Preview on productivity tasks while costing 1/13 as much per task. Following a surge in demand for the model, Tencent's WorkBuddy platform expanded its inference cluster tenfold."
---

# Tencent Releases Open Weights for Hy4 Flash Multimodal AI Model

> [Open the canonical story](<https://aidr.today/d369c4d2?lang=en>)

**Published:** 2026-11-18T16:52:00.000Z
**Category:** Models
**Topics:** open\-source

## Summary

Tencent released the open weights for its Hy4 Flash multimodal AI model, including FP8 and Base versions\. The model uses a mixture\-of\-experts architecture with 224B total parameters and 12B active parameters per token, supporting a 1M token context window\. The weights are distributed under the Tencent Hy License\. The release coincides with day\-0 support from the vLLM and SGLang runtimes\. Tencent claims Hy4 Flash slightly outperforms its larger 770B parameter Hy4 Preview on productivity tasks while costing 1/13 as much per task\. Following a surge in demand for the model, Tencent's WorkBuddy platform expanded its inference cluster tenfold\.

## Sources

- [Story source](<https://marketbrief.now/ai/tencent-releases-open-weights-for-hy4-flash-multimodal-ai-model-e35843f3>)
- [Story source](<https://x.com/TencentHunyuan/status/2122912575302119554>)
- [Story source](<https://x.com/TencentHunyuan/status/2122217770314301469>)
- [Story source](<https://x.com/vllm_project/status/2122914790377164886>)
- [Story source](<https://x.com/vllm_project/status/2122917809948852559>)
- [Supporting source](<https://x.com/Sino_Market/status/2122979364631720140>)
- [Supporting source](<https://x.com/himanshustwts/status/2122319244042047969>)
- [Supporting source](<https://x.com/martinvars/status/2122399510882074750>)

