Tencent Open-Sources Hunyuan HY3: 295B MoE with Switchable Reasoning
AdinaYakup · x · 2026-07-06
Tencent has released and open-sourced the Hunyuan HY3 model, featuring a 295B parameter / 21B activation Mixture-of-Experts (MoE) architecture. It supports a 256K ultra-long context, offers an FP8 quantized version, and is licensed under Apache 2.0.
The model supports switchable reasoning modes (no thinking / low intensity / high intensity), reduces hallucination rates by 50% compared to its predecessor, and demonstrates stable performance across mainstream agentic frameworks.
Targeting developer and enterprise deployment scenarios with a lenient open-source license, it serves as a major addition to the domestic open-source LLM landscape.
Related event: Tencent Open-Sources Hunyuan Hy3 for Agentic Workloads(26 posts)→
More from Models
- Developer Building a Unified Leaderboard of All Model Benchmark Scores — airesearch12 · 2026-09-11
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11
- GPT-5.6 writes well but is instantly forgettable, user complains — BasedRaddka · 2026-09-11
- Opus Refuses Protein Research Codebase Over 'Safety' Concerns, Dev Considers Rolling His Own — josephdviviano · 2026-09-11
- User Hails Unconfirmed 'DeepSeek 4.1 Flash' as an Inflection Point in LLMs — himanshustwts · 2026-09-11
- Terminal Bench v4: GLM-5.3 Leads at 41.9%, Kimi-K3 Underwhelms at 12.6% — Ok_Warning2146 · 2026-09-11