Skip to main content
Back to News
Tencent Hunyuan Releases Hy4 Preview as an Open Model for Productivity Work
Product
2 min read

Tencent Hunyuan Releases Hy4 Preview as an Open Model for Productivity Work

The AMW Read

Tencent's open Hy4 Preview materially advances its position in the open-model field by combining large-model scale, long context, product distribution, and API availability.
NoveltySignificance
Foundation Models · Player MapScaling Laws

Tencent Hunyuan Releases Hy4 Preview as an Open Model for Productivity Work

Tencent Hunyuan released and open-sourced Hy4 Preview on August 28. The model uses 770 billion total parameters with 49 billion activated parameters and supports a context window above one million tokens. Tencent positions it for software engineering, office analysis, game development, and scientific research. In an internal blind evaluation involving 163 experts and 203 engineering tasks, Tencent said Hy4 Preview scored 2.99 out of 4, compared with 2.92 for GLM 5.3 and 2.94 for Kimi K3.

The release matters less as a standalone benchmark claim than as another major Chinese platform company pairing open model weights with its own distribution surfaces. Tencent launched Hy4 Preview simultaneously in WorkBuddy, CodeBuddy, Yuanbao, and ima, while also offering API access through Tencent Cloud Tokenhub and OpenRouter. Its stated pricing is RMB 6 per million input tokens, RMB 18 per million output tokens, and RMB 0.3 per million cached tokens. That combination gives Tencent a route to gather usage feedback from productivity products while competing for external developers on cost and availability.

Builders should test Hy4 Preview on long-context and multi-file workflows that resemble their actual deployments, rather than relying on Tencent's internal scorecard. The two-week free-access period in WorkBuddy and CodeBuddy can lower the cost of evaluation, while the API channels make comparison against other open models more practical. Investors should watch whether Tencent can turn product distribution and rapid iteration into externally validated adoption, especially in coding and office workflows where model quality, integration, and inference economics are tightly linked.

#Tencent #Hunyuan #OpenSourceAI #FoundationModels #AICoding

#Tencent Hunyuan#Hy4 Preview#open-source models#AI productivity

How This Connects

Based on Foundation Models · Player Map

  1. 4h agoMistral Raises €3B Series D at €21B+ Valuation, Led by Samsung ElectronicsMistral
  2. 1d agoAnthropic assembles $517 billion in compute commitments over 11 monthsAnthropic
  3. 1w agoTencent Hunyuan Releases Hy4 Preview as an Open Model for Productivity Work · THIS ARTICLE
  4. 3w agoAlibaba's Qwen team has open-sourced Qwen3.8-27B, a 27-billion-parameter multimodal model designed f...Qwen
  5. 3w agoAlibaba has released Qwen 3.8 27B, an Apache 2.0-licensed open-weight dense model with 27 billion pa...Alibaba Qwen 3.8 27B launch
  6. 1mo agoMoonshot AI launches Kimi K3, a 2.8 trillion-parameter open-weight model, claiming performance near US frontier labsMoonshot AI

Related News

More news from Tencent (Hunyuan AI Model Family)

Stay updated with the latest news and announcements from Tencent (Hunyuan AI Model Family).

View all Tencent (Hunyuan AI Model Family) news

Discover AI Startups

Explore 5,000+ AI companies with VC-grade analysis, funding data, and investment insights.

Explore Dashboard