Alibaba/Mixture of Experts

Qwen3-Coder-Next

Name: Qwen3-Coder-Next
Author: Alibaba

Qwen3-Coder-Next — 80B MoE focused on coding and tool use.

chatcodingreasoningtool_useThinkingTool Use

80B

Parameters (3B active)

256K

Context length

Benchmarks

Quantizations

1.2M

HF downloads

Architecture

MoE

Released

2026-01-30

Layers

KV Heads

Head Dim

256

Family

qwen

Quantization Options

Quant	Bits	VRAM	Quality
IQ2_M	2.93	29.8 GB	low
Q2_K	3.16	32.1 GB	low
IQ3_XXS	3.25	33.0 GB	low
IQ3_XS	3.5	35.5 GB	low
Q3_K_S	3.64	36.9 GB	low
IQ3_M	3.76	38.1 GB	low
Q3_K_M	4	40.5 GB	low
Q3_K_L	4.3	43.5 GB	moderate
IQ4_XS	4.46	45.1 GB	moderate
Q4_K_S	4.67	47.2 GB	moderate
Q4_K_M	4.89	49.4 GB	good
Q5_K_S	5.57	56.2 GB	good
Q5_K_M	5.7	57.5 GB	good
Q6_K	6.56	66.1 GB	excellent
Q8_0	8.5	85.5 GB	lossless
FP16	16	160.5 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 QWEN3-CODER-NEXT 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 (17)

IFEval85.9

τ²-Bench79.5

GPQA Diamond73.7

SWE-bench70.6

BBH60.5

MATH60.1

MMLU-PRO50.4

AA Long Context40.0

IFBench35.2

BigCodeBench33.2

SciCode32.3

AA Intelligence28.3

AA Coding22.9

GPQA19.4

Terminal-Bench18.2

MUSR12.3

HLE9.3

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

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

$ollama run qwen3-coder-next:q4_k_m

Downloads and runs automatically. Add --verbose for speed stats.

▸ 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 Ollama Library GGUF Downloads Build Hardware