Tencent Hunyuan Hy3 Officially Released with Day-One Native vLLM Support
vllm_project · x · 2026-07-06
Tencent Hunyuan has released the official version of Hy3 (succeeding the Hy3 Preview). It natively supports the vLLM inference framework from day one and is equipped with tool call parsing, reasoning parsing, and MTP speculative decoding, verified on both NVIDIA and AMD hardware.
Hy3 is a MoE model designed for agentic workflows, coding, and long-horizon reasoning, released under the Apache 2.0 license. It features 295B total parameters with 21B activated, 192 experts with top-8 routing, GQA attention, and a 256K context window. It also includes a 3.8B parameter MTP speculative decoding layer, with both BF16 and FP8 versions available.
Related event: Tencent Open-Sources Hunyuan Hy3 for Agentic Workloads(26 posts)→
More from Infra
- $5M Seed Round Can't Afford One DeepSeek Training Run: GPU-less AI Startup's Three Paths — FinanceYF5 · 2026-10-26
- HBM demand shifts: low-stack inference plans shelved, Rubin to go 12-high only — BenBajarin · 2026-10-26
- CircuitHub unveils The Grid, a factory-scale robot making PCBs in the US twice as fast as Shenzhen — garrytan · 2026-10-26
- Ubicloud publishes "Fixing Tool-Call Accuracy in vLLM" — petrusenko_max · 2026-10-26
- Meta approves nonvoting Class C shares as 2026 capex guidance hits $135B–$145B — thione · 2026-10-26
- Unsloth Ships MiniMax-M3.1-Flash Quants: 2-bit Retains 91.5% Accuracy, Runs in 150GB RAM — danielhanchen · 2026-10-26