Arrakis / Models / Hy3

Hy3 GPU requirements

Hy3 has 299B parameters. To serve it, plan for about 717 GB of GPU memory in 16-bit, 359 GB in 8-bit or 179 GB in 4-bit, including about 20% for the KV cache and activations.

It is a mixture of experts: 192 experts, 8 of them active for each token. All of them stay in GPU memory, so plan memory for all 299B parameters.

Deploy this model Check another model

What it takes to serve it

GPU memory to run it for inference, with the smallest GPU setup that holds it and the lowest price on Arrakis right now.

Precision	Memory and the smallest setup
16-bit	717 GB: 4 × B200, from $25.32 per hour for all 4
8-bit	359 GB: 2 × B200, from $12.66 per hour for all 2
4-bit	179 GB: one B200, from $6.33 per GPU-hour

What it takes to fine-tune it

QLoRA trains adapters on a 4-bit frozen model, LoRA on a 16-bit one, and full fine-tuning updates every weight. Figures are before activations, which grow with batch size and sequence length.

Method	Memory and the smallest setup
QLoRA	209 GB: one B300, from $4.60 per GPU-hour
LoRA	717 GB: 4 × B200, from $25.32 per hour for all 4
Full fine-tuning	4,781 GB: more than one 8-GPU machine

About the model

From its card and files on Hugging Face. Memory figures here follow the GPU memory guide.

Parameters: 299B
Experts: 192
Experts for each token: 8
Context window (tokens): 262,144
Published in: BF16
License: apache-2.0
Released: 2026-07-02
Weights: tencent/Hy3

Deploy it on Arrakis

Open Models in the console, choose vLLM, SGLang or llama.cpp, and rent a GPU it fits on. You get an OpenAI-compatible endpoint with its own API key, and root SSH to the machine. Or run it yourself on any machine with enough GPU memory:

vLLM

vllm serve tencent/Hy3 --tensor-parallel-size 4

Deploy this model How deploying works

Questions

Live prices for every GPU are on the GPU prices page.

How much GPU memory does Hy3 need?

To serve it, about 717 GB in 16-bit, 359 GB in 8-bit, 179 GB in 4-bit, including about 20% for the KV cache and activations. To fine-tune it, about 209 GB with QLoRA, 717 GB with LoRA and 4,781 GB for full fine-tuning with Adam, before activations.

Can I run Hy3 on one GPU?

Yes, quantized. In 4-bit it needs about 179 GB, which fits on one B200. In 16-bit it needs 4 B200 GPUs.

What does it cost to run Hy3?

In 16-bit it needs 4 B200 GPUs; in 4-bit, one B200. Each setup on this page shows the lowest price per GPU-hour on Arrakis right now, and you pay by the hour from a prepaid balance.

How do I deploy Hy3 on Arrakis?

Open Models in the Arrakis console, paste tencent/Hy3, choose vLLM, SGLang or llama.cpp, and rent a GPU it fits on. You get an OpenAI-compatible endpoint with its own API key, and root SSH to the machine.

More models

Every model, and a check for any other.