Tencent Hunyuan Hy4 Preview Cuts Token Usage: what are the key facts?

Tencent's Hunyuan team and WorkBuddy have fully deployed an optimized version of the Hy4 preview model.

Tencent Hunyuan Hy4 Preview Cuts Token Usage: what are the key facts?

The update specifically addresses excessive self-verification and long reasoning chains in complex tasks.

Tencent Hunyuan Hy4 Preview Cuts Token Usage: what are the key facts?

Post-optimization, the model significantly reduces task rounds and token consumption while maintaining performance.

Tencent Hunyuan Hy4 Preview Cuts Token Usage: what are the key facts?

Released on August 28, the model features 77 billion total parameters with 4.9 billion active.

Physix Frontier · News Briefing Card (enbrief · Sep 8, 2026)

Tencent Hunyuan Hy4 Preview Cuts Token Usage

KEY FACTS

Tencent's Hunyuan team and WorkBuddy have fully deployed an optimized version of the Hy4 preview model.
The update specifically addresses excessive self-verification and long reasoning chains in complex tasks.
Post-optimization, the model significantly reduces task rounds and token consumption while maintaining performance.
Released on August 28, the model features 77 billion total parameters with 4.9 billion active.

KEY DATA

770 BTotal Parameters

49 BActive Parameters

1MContext Length

PHYSIX OBSERVATION

This update targets a critical pain point in LLM applications: prohibitive inference costs. By algorithmically suppressing inefficient internal processing to compress tokens without sacrificing performance, Tencent lowers the barrier for enterprise deployment. This signals a shift in open-source competition from mere parameter scaling to balancing inference efficiency with economic viability, benefiting real-world implementation scenarios.

Source: enbrief original report ↗