Tencent/Mixture of Experts

THunyuan Hy3-preview

Name: Hunyuan Hy3-preview
Author: Tencent

🖥️ Official Website  |  

chatcodingreasoningmultilingualmathtool_use

295B

Parameters (21B active)

256K

Context length

Benchmarks

Quantizations

Architecture

MoE

Released

2026-04-23

Layers

KV Heads

Head Dim

128

Family

hunyuan

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_XXS	2.38	88.3 GB	low
IQ2_M	2.93	108.5 GB	low
Q2_K	3.16	117.0 GB	low
IQ3_XXS	3.25	120.3 GB	low
IQ3_XS	3.5	129.6 GB	low
Q3_K_S	3.64	134.7 GB	low
IQ3_M	3.76	139.1 GB	low
Q3_K_M	4	148.0 GB	low
Q3_K_L	4.3	159.1 GB	moderate
IQ4_XS	4.46	165.0 GB	moderate
Q4_K_S	4.67	172.7 GB	moderate
Q4_K_M	4.89	180.8 GB	good
Q5_K_S	5.57	205.9 GB	good
Q5_K_M	5.7	210.7 GB	good
Q6_K	6.56	242.4 GB	excellent
Q8_0	8.5	313.9 GB	lossless
FP16	16	590.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN HUNYUAN HY3-PREVIEW NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (8)

GPQA Diamond87.2

MBPP78.7

MATH76.3

SWE-bench74.4

MMLU-PRO65.8

Terminal-Bench54.4

LiveCodeBench34.9

HLE30.0

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run hunyuan:295b-q4_K_M

Tag may need adjustment — check ollama.com/library/hunyuan for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware