Newsletter Subscribe
Enter your email address below and subscribe to our newsletter
[forminator_form id="25163"]

openrouter+1tencent+1forbes+1Tencent's flagship large language model Hy3 has seen its total API call volume grow more than 68 times compared to its predecessor Hy2, just one week after its official launch on July 6, according to the company's Tencent Hunyuan team. The model has also claimed the top spot on OpenRouter's global weekly leaderboard for model usage, processing 6.13 trillion tokens.openrouter+3
Among users of WorkBuddy — Tencent's enterprise productivity assistant — who self-select their AI model, 60% have chosen Hy3, the company said in a post on X. The model was integrated at launch into several Tencent products including WorkBuddy, CodeBuddy, Yuanbao, Marvis, and ima.tencent+3
Hy3 is built on a Mixture-of-Experts architecture with 295 billion total parameters and 21 billion active parameters, supporting a context window of up to 256K tokens. Tencent has positioned the model not as a benchmark champion but as one optimized for real-world AI agent workflows, coding assistants, and enterprise productivity tasks.openrouter+3
The official Hy3 release builds on the Hy3 Preview, which launched on April 23 and itself topped OpenRouter's daily usage rankings within days of release. Two weeks after that preview launch, call volume had already exceeded Hy2 by more than 10 times, while token usage in agent applications rose 16.5 times. The full Hy3 release has accelerated that trajectory considerably.news.futunn+2
Tencent priced Hy3 at approximately RMB 1 per million input tokens and RMB 4 per million output tokens, and open-sourced it under the Apache 2.0 license on Hugging Face. Free access is available on OpenRouter until July 21.technode+2
The surge places Tencent's model ahead of Xiaomi's MiMo-V2.5 and DeepSeek's V4 Flash on the OpenRouter leaderboard, underscoring how Chinese open-weight models continue to gain traction among global developers. Forbes reported that Tencent is marketing Hy3 as a model built for reliability and cost efficiency rather than raw scale, a strategy that appears to be resonating with developers seeking production-ready AI capabilities.forbes+1