Tencent open-weights Hunyuan Hy4 flash: 224B MoE with 12B active params and 1M context
TencentHunyuan · x · 2026-11-18
Tencent's Hunyuan team has released Hy4 flash as open weights, including FP8 and Base variants, available on Hugging Face and GitHub.
- Specs: sparse MoE with 224B total / 12B active parameters and 1M token context
- Natively multimodal, with dedicated image and video tokens plus structured thinking/tool-call markers visible in the chat template
- Positioned as a fast, efficient model for everyday productivity
- Released under the Tencent Hy License; the full Hy4 model is teased as coming next
More from Models
- Counterintuitive: capping Qwen3.8-27B's thinking budget lifts GPQA score from 115 to 160 — AIFlow_ML · 2026-11-18
- Anthropic's own summarizer refuses its requests, exposing layers of model guardrails — zetalyrae · 2026-11-18
- PenaltyBench: Gemini 4 Carbon beats VAR referees at calling penalties, 63.3 vs 62.3 — ziqiao_ma · 2026-11-18
- Tencent quietly lists Hy4-flash model page on Hugging Face — BreakfastFriendly728 · 2026-11-18
- Claude models 'discover' body parts only when touched, says researcher — repligate · 2026-11-18
- GPT-6.2 Sol hands-on: Fast mode finishes tasks noticeably quicker, says KOL — op7418 · 2026-11-18