⚡ Uncle Cat AI Radar
IndustryModelsAgents

Tencent takes Hy3 global with free WorkBuddy access

Tencent opened its 295B-parameter Hy3 model to overseas users through WorkBuddy's international edition, cloud APIs and design tools, free until August 31.

Tencent said on Wednesday that Hy3, the flagship Hunyuan model it launched domestically on July 6, is now available worldwide across three surfaces: the international edition of its WorkBuddy agent, the Miora design product, and TokenHub on Tencent Cloud. Access through WorkBuddy is free for individual and enterprise users until August 31, Pacific time.

The offer

Hy3 is a mixture-of-experts model with 295 billion total parameters and 21 billion active, a 256K context window, and what Tencent describes as hybrid fast-and-slow thinking — routing between quick responses and extended reasoning. The company positions it as matching flagship models several times its size on parameter count while costing far less to serve; on OpenRouter it is listed at roughly $0.13 per million input tokens and $0.53 per million output tokens.

Tencent leaned on adoption metrics rather than benchmarks. It says Hy3 has drawn 68 times the API call volume of its predecessor and reached the top of OpenRouter's global usage leaderboard within a week of release. For WorkBuddy specifically — an agent with concurrent multi-agent execution and more than 100 built-in skills, handling research, data analysis, document and presentation generation — internal evaluations put task success above 90 percent and average completion time 34 percent lower than the previous Hunyuan generation. Serving runs across 66 availability zones in 23 regions.

Why it matters

The interesting part is not the model but the distribution strategy. Chinese labs have historically globalized through open weights or bare API endpoints; Tencent is instead exporting a finished agent product with a month of free access, aimed at the same knowledge-work budgets that OpenAI, Anthropic and Google are chasing with enterprise agents. That makes token economics the wedge: a 21-billion-active-parameter model priced near a tenth of Western flagships can absorb the many-call, long-horizon workloads agents generate. Whether overseas enterprises will route work through a Chinese-hosted model remains the open question, and data-residency scrutiny in Europe and the US is the obvious constraint. But the free window is a deliberate test of how far price and packaging can move that decision.

Sources