FitMyLLM
Tencent/Mixture of Experts

TencentHunyuan-Large

Hunyuan-Large from Tencent: MoE, 389B total and 52B active, 128K context. Tagged for coding, reasoning, multilingual, math.

chatcodingreasoningmultilingualmath
389B
Parameters (52B active)
128K
Context length
7
Benchmarks
17
Quantizations
0
Architecture
MoE
Released
2024-11-05
Layers
64
KV Heads
8
Head Dim
128
Family
hunyuan

Quantization Options

Context length:
QuantBitsVRAM @ 16KQuality
IQ2_XXS2.38
119.2 GB
116.2 + 3.0 KV
low
IQ2_M2.93
146.0 GB
143.0 + 3.0 KV
low
Q2_K3.16
157.1 GB
154.1 + 3.0 KV
low
IQ3_XXS3.25
161.5 GB
158.5 + 3.0 KV
low
IQ3_XS3.5
173.7 GB
170.7 + 3.0 KV
low
Q3_K_S3.64
180.5 GB
177.5 + 3.0 KV
low
IQ3_M3.76
186.3 GB
183.3 + 3.0 KV
low
Q3_K_M4
198.0 GB
195.0 + 3.0 KV
low
Q3_K_L4.3
212.6 GB
209.6 + 3.0 KV
moderate
IQ4_XS4.46
220.4 GB
217.4 + 3.0 KV
moderate
Q4_K_S4.67
230.6 GB
227.6 + 3.0 KV
moderate
Q4_K_M4.89
241.3 GB
238.3 + 3.0 KV
good
Q5_K_S5.57
274.3 GB
271.3 + 3.0 KV
good
Q5_K_M5.7
280.7 GB
277.7 + 3.0 KV
good
Q6_K6.56
322.5 GB
319.5 + 3.0 KV
excellent
Q8_08.5
416.8 GB
413.8 + 3.0 KV
lossless
FP1616
781.5 GB
778.5 + 3.0 KV
lossless

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

Too big for a single GPU — plan a multi-GPU deployment
Even the lightest quant needs ~119 GB. Size GPUs, replicas, TCO and scaling for a production setup. Open in Enterprise →
Ready to run this?Rent by the hour

Rent a GPU and run Hunyuan-Large now

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

Community Ratings

Loading ratings...

Benchmarks (7)

Run this model

Easiest way to get started·Beginners
Docs ↗
curl -fsSL https://ollama.com/install.sh | sh
$# Hunyuan-Large is not in Ollama's library, and we know no GGUF build of it

Search Hugging Face for a GGUF of it (https://huggingface.co/models?sort=trending&search=Tencent-Hunyuan-Large+GGUF), or use the vLLM tab with the original weights.

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

GPUs that can run this model

At Q4_K_M quantization. Sorted by minimum VRAM.

AMD Radeon Instinct MI325X
288 GB VRAM • 10300 GB/s
AMD
$20000
AMD Radeon Instinct MI350X
288 GB VRAM • 8190 GB/s
AMD
$25000
AMD Radeon Instinct MI355X
288 GB VRAM • 8190 GB/s
AMD
$30000
Apple M4 Ultra (384GB)
384 GB VRAM • 1092 GB/s
APPLE
$9999
Apple M5 Ultra (384GB)
384 GB VRAM • 1228 GB/s
APPLE

Find the best GPU for Hunyuan-Large

Build Hardware for Hunyuan-Large
Spec sheet

Hunyuan-Large — 389B MoE.

Specifications
Parameters
389B (52B active)
Architecture
Mixture of Experts
Context Length
128K tokens
Capabilities
chat, coding, reasoning, multilingual, math
Release Date
2024-11-05
Provider
Tencent
Family
hunyuan
VRAM requirements
QuantBPWVRAMQuality
IQ2_XXS2.38116.2 GB65%
IQ2_M2.93143.0 GB75%
Q2_K3.16154.1 GB78%
IQ3_XXS3.25158.5 GB82%
IQ3_XS3.5170.7 GB84%
Q3_K_S3.64177.5 GB85%
IQ3_M3.76183.3 GB86%
Q3_K_M4195.0 GB88%
Q3_K_L4.3209.6 GB90%
IQ4_XS4.46217.4 GB92%
Q4_K_S4.67227.6 GB93%
Q4_K_M4.89238.3 GB94%
Q5_K_S5.57271.3 GB96%
Q5_K_M5.7277.7 GB96%
Q6_K6.56319.5 GB97%
Q8_08.5413.8 GB100%
FP1616778.5 GB100%
Benchmark scores
HumanEval90.0
MMLU-PRO60.2
MATH77.4
IFEval85.0
BBH89.5
MBPP72.6
GPQA Diamond42.4
Compatible GPUs
5 @ Q4_K_M