Skip to main content
Watch: I Built an AI App Without Leaving Databricks

Tencent - Mixture of experts - verified

Run Hy3 locally

Hy3 is a 298.8B mixture-of-experts (21B active) model from Tencent with a 262K-token context window, released under Apache 2.0 (commercial use is allowed); the smallest listed download is 182.3 GB, the Q4_K_M GGUF build from AngelSlim/Hy3-GGUF.

Official weights: tencent/Hy3

Total params
298.8B
Active params
21B
Context
262K
Smallest download
182.3 GB

Fits in

Smallest download at most 85% of memory

Downloads

The official weights and every quantization listed for this model, with file sizes from Hugging Face.

Hy3 downloads and file sizes
DownloadFormatSizeRepo
Official weightsBF16597.6 GBtencent/Hy3
Q4_K_MSmallest listedGGUF182.3 GBAngelSlim/Hy3-GGUF

GB is decimal: 1 GB = 10^9 bytes. Add memory for the context and the operating system on top of the file size.

Run it

Copied verbatim from the page linked under each command. Server commands from model cards often assume a multi-GPU node; check the flags before you run them.

vLLM

Terminal
# Switch to trtllm backend to work-around mnnvl workspace size issue.
export VLLM_FLASHINFER_ALLREDUCE_BACKEND=trtllm
vllm serve tencent/Hy3 \
  --tensor-parallel-size 8 \
  --speculative-config.method mtp \
  --speculative-config.num_speculative_tokens 2 \
  --tool-call-parser hy_v3 \
  --reasoning-parser hy_v3 \
  --enable-auto-tool-choice \
  --port 8000 \
  --served-model-name hy3

Source: huggingface.co/tencent/Hy3

SGLang

Terminal
python3 -m sglang.launch_server \
  --model tencent/Hy3 \
  --tp-size 8 \
  --tool-call-parser hunyuan \
  --reasoning-parser hunyuan \
  --speculative-num-steps 2 \
  --speculative-eagle-topk 1 \
  --speculative-num-draft-tokens 3 \
  --speculative-algorithm EAGLE \
  --port 8000 \
  --served-model-name hy3

Source: huggingface.co/tencent/Hy3

Not sure which runtime to use? Compare local LLM runtimes by platform, API and features.

License

License
Apache 2.0apache-2.0
Commercial use
Commercial use: YesCommercial use is allowed.
Gated download
No. The files download directly.

Worth knowing

Hy3 is a 192-expert top-8 MoE with an MTP layer for speculative decoding and a reasoning_effort setting of no_think, low or high.

Released (Hugging Face repo creation date).

API pricing and specsHosted prices per 1M tokens for Hy3, if you would rather not run it yourself.

FAQ

How big is the Hy3 download?

The official BF16 weights in tencent/Hy3 are 597.6 GB. The smallest listed download is 182.3 GB, the Q4_K_M GGUF build from AngelSlim/Hy3-GGUF. Plan for more memory than the file size: the context (KV cache) and the operating system need room too.

Can I use Hy3 commercially?

Hy3 is released under Apache 2.0, and commercial use is allowed. The license text is at https://www.apache.org/licenses/LICENSE-2.0.

How much memory do I need for Hy3?

By this directory's rule of thumb it fits in 256 GB and up: the smallest listed download, 182.3 GB, is at most 85% of 256 GB, which leaves room for the context and the operating system.

What is the context length of Hy3?

262,144 tokens. A longer context needs more memory for the KV cache, so you can run with a shorter one when memory is tight.

Sources

Last verified from the pages below.

Newsletter

New open weights, every week

One email per week. New model releases, open-weight drops and what they mean for what you build.

Keep exploring

Best local AI models

Our curated picks by hardware tier, with benchmarks and the reasoning.

Model API pricing and specs

First-party prices per 1M tokens when you would rather call an API.

AI benchmark leaderboards

Sourced scores, for when you need to compare quality.

Setup guides

Point your coding tools at the model you picked.