Model index
Tencent Hy·Hunyuan / Hy

Hy4 preview

Tencent Hy’s next-generation 770B MoE flagship preview with 49B active parameters and a 1M context for coding, office analysis, game development, and research.

PreviewLanguageReasoningCodingAgentOpen weights
MODEL NUMBER186
First public date
2026-08-28
Organization
Tencent Hy
Family
Hunyuan / Hy
01

Specifications

Only published specifications are shown.

Total parameters
770B↗
Active parameters
49B↗
Architecture
MoE↗
Parameter scope
Backbone parameters and MTP components are separate scopes; platform tensor totals are not treated as backbone parameters.↗
Context
1,000,000 tokens
Input modalities
Text
Output modalities
Text
Open weights
Yes
License
Apache-2.0
Status
Preview
02

Overview

Release dates

— Tencent released and open-sourced Hy4 preview. ↗

Why it mattered

Hy4 preview expands model scale, context, and training data, bringing Gated DSA, IndexCache, iHC, and native MTP into Hy’s open-model line.

03

Specifications alongside predecessor

Compared with Hy3 ↗

PreviousHy3295B · 256K ctx
CurrentHy4 preview770B · 1M ctx
04

Architecture & capabilities

Attention

Gated DSAIndexCache

Model capabilities

Deep reasoningCodingTool useLong-horizon agentScientific research

Product features

Professional work
05

Release history

Open weights

Tencent released and open-sourced Hy4 preview.

View source
06

Related models

Select a card to view the related model.

07

Sources

Suggest a correction

Copy this template, add the field, proposed value, and primary source, and share it with the archive maintainer.