Tencent releases Hy-Embodied-VLM-1.0 Mixture-of-Experts model
July 15, 2026
Tencent's Hy-Embodied-VLM-1.0 utilizes a Mixture-of-Experts architecture for embodied agents, activating only 3B parameters per token. The model achieves state-of-the-art performance on 19 of 38 evaluated benchmarks.
HOW THIS AFFECTS YOU
●
builderThis provides a more efficient VLM option for robotics and embodied AI agents due to low active parameter counts.
●
researcherThe MoE approach achieves SOTA results while maintaining high inference efficiency.