Tencent Hunyuan Hy4 preview: 770B total/49B active, 1M context, Apache 2.0, day-0 vLLM
aftahi_ai · x · 2026-09-05
Tencent's Hunyuan Hy4 preview packs 770B total / 49B active params, native 1M context, Apache 2.0 licensing, day-0 vLLM and SGLang support, and official FP8 weights. Mixed per-layer quantization cuts memory from 1.5TB to 214GiB with reported accuracy barely moving vs BF16. A demo shows it turning a simple prompt into a playable game — a 770B-class model with a far lower deployment barrier.
Related event: Tencent Open-Sources Hunyuan Hy4 Preview with 770B Params and 1M Context(3 posts)→
More from Infra
- Jensen Huang visits Astana, joins Kazakhstan's AI minister to discuss nation's AI future — nvidia · 2026-10-23
- FP16 vs FP32 matmul on RTX 4060: 3.5x speedup at large sizes, slower at 128x128 — cneuralnetwork · 2026-10-23
- BaseQuant V1 quants for agentic Qwen3.8-27B hit 17-34% lower KL than Unsloth — nightowlsleeping · 2026-10-23
- PhD students in some countries get B200s on demand, unlike their US peers — HildeKuehne · 2026-10-23
- Nvidia's Berlin GTC spotlights humanoid robots and Vera Rubin AI supercomputing — emmanuelvivier · 2026-10-23
- Lenovo adds 5 AI Library use cases, claims 1-week deployment and up to 70% lower token cost — shashib · 2026-10-23