VQV Signal

OPEN SOURCE WATCH TECHNICAL

Tencent Hunyuan releases open weights for Hy4 multimodal mixture-of-experts model

According to a post on X by @TencentHunyuan, Tencent Hunyuan has released open weights for Hy4, its natively multimodal mixture-of-experts flagship with 770B total parameters, 49B active parameters and a 1M-token context. The release includes Hy4, Hy4-FP8 and Hy4-base builds, plus GGUF builds and a System Card covering safety evaluations. The post says vLLM announced day-0 support and that SGLang is also supported. The source text is truncated, so further details are not available here.

Source: Scobleizer News Pulse (X) · x.com Found via Robert Scoble’s Scobleizer News Pulse · a post on X, not independently verified Published 2026-11-20T06:08:00+00:00 Detected 2026-11-20T19:01:59+00:00
View original source

According to a post on X by @TencentHunyuan, Tencent Hunyuan has released open weights for Hy4, its natively multimodal mixture-of-experts flagship with 770B total parameters, 49B active parameters and a 1M-token context. The release includes Hy4, Hy4-FP8 and Hy4-base builds, plus GGUF builds and a System Card covering safety evaluations. The post says vLLM announced day-0 support and that SGLang is also supported. The source text is truncated, so further details are not available here.

AI-assisted summary based on the listed source.

Tencent Hunyuan has released open weights for Hy4, its natively multimodal mixture-of-experts flagship (770B total, 49B active parameters, 1M-token context), as Hy4, Hy4-FP8 and Hy4-base plus GGUF builds, alongside a System Card with its safety evaluations. vLLM announced day-0 support and Tencent says SGLang is...

Developers and businesses that work with open-weight models can now examine Hy4 directly, including its FP8 and GGUF builds and the safety evaluations in the System Card. The stated support from vLLM and SGLang suggests the model can be run with common inference tools, though performance and license terms are not covered in the material provided. Watch for independent testing and the full release details to see how the model performs in practice.

Signal Strength 95% Technical label WATCH Public Interest 67 Category OPEN SOURCE Reader Depth TECHNICAL

Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.

Public Interest components
Recognizable Entity Score 51 Practical Impact Score 98 Novelty Interest Score 94 Consequence Score 34 Curiosity Score 32 Shareability Score 74

VQV surfaced this signal because it is recent, relevant to Open Source LLMs, connected to Scobleizer News Pulse (X).