back to local model checker

Qwen 3 Coder Next 80B-A3B

MoETools

Alibaba · 80B (3B active) · Mixture of Experts

Ultra-efficient agentic coding MoE optimized for tool-calling coding agents

Apache 2.0
Downloads

0

Likes

0

Released

2026-02

Context

256K tokens

Use cases

code
Total experts512
Active experts10
Active params3.0B

Your hardware

16 GB Reference Desktop

Quant
Q4_K_M
Memory
41.5 GB
Speed
--
Fit
Too heavy

Quantization options

What is quantization?
QuantBitsVRAMDiskQualityStatus
Q2_K226.1 GB24.4 GBlowBarely runs · ~2 t/s
Q3_K_M336.4 GB34.2 GBmoderateBarely runs · ~1 t/s
Q4_K_M441.5 GB39.1 GBgoodToo heavy · ~1 t/s
Q5_K_M551.7 GB48.9 GBgoodToo heavy · ~1 t/s
Q6_K662 GB58.7 GBexcellentToo heavy · ~1 t/s
Q8_0882.5 GB78.2 GBexcellentToo heavy · ~1 t/s
F1616164.4 GB156.5 GBlosslessToo heavy · ~1 t/s

About this model

Qwen 3 Coder Next 80B-A3B (Qwen3 coder next) is a 80B sparse mixture-of-experts (MoE) model activating 3B active per token from Alibaba in the Qwen family, designed primarily for code workflows. Supports structured tool calling; use Q4_K_M or higher when strict JSON schema adherence matters. Supports up to 256K context tokens; start at 4K–8K with Q8_0 KV cache on tight GPUs.

Hardware planning baselines start at 44.7 GB minimum memory and 74.5 GB recommended memory. At Q4_K_M, weights plus baseline runtime overhead use roughly 41.5 GB of RAM/VRAM and 39.1 GB of disk space. At 80B, Qwen 3 Coder Next 80B-A3B belongs firmly in the workstation and multi-GPU tier. Running it comfortably requires 74.5 GB of total memory across dual RTX GPUs, high-memory Apple Silicon (Mac Studio Ultra/Max), or MoE hybrid CPU-RAM offloading—mobile devices and standard 16 GB laptops are too constrained for practical speeds.

Deploy Qwen 3 Coder Next 80B-A3B locally via Ollama (`qwen3-coder-next`), Ollama with Flash Attention enabled. Under Apache License 2.0, review official weight cards before commercial redistribution.

Highlights

  • Local Fit Baseline44.7 GB minimum and 74.5 GB recommended RAM/VRAM.
  • Primary QuantQ4_K_M targets 41.5 GB memory and 39.1 GB disk.
  • Context & Capabilities256K max context · Sparse MoE · Tool calling.

System requirements by device

Memory planning targets 44.7 GB minimum RAM/VRAM and 74.5 GB recommended memory. The baseline quant is Q4_K_M, estimated at 41.5 GB memory and 39.1 GB disk. Actual runtime usage shifts with context length, KV cache precision, and OS background overhead.

DeviceNotes
Mac (Apple silicon)Requires a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra with enough unified wired budget for 41.5 GB.
Windows / Linux GPUUse dual/multi-GPU CUDA splitting (--tensor-split) or MoE CPU expert offloading (-ot) paired with 32–128 GB DDR5 system RAM.
Phones & Compact Edge BoardsNot practical on phones (44.7 GB min RAM exceeds Android LMK and iOS Jetsam limits). Pick a 2B–4B model on mobile.

Start with a short session and leave memory for the operating system. A file fitting on disk does not guarantee that it will load in memory.

Estimated memory by quantization

These are rough estimates, not tested minimum requirements. Check the actual file and runtime documentation; context, offloading and operating-system overhead can change the memory needed.

Quant TierEstimated memoryDisk SizeQuality
Q2_K26.1 GB24.4 GBlow
Q3_K_M36.4 GB34.2 GBmoderate
Q4_K_M41.5 GB39.1 GBgood
Q5_K_M51.7 GB48.9 GBgood
Q6_K62.0 GB58.7 GBexcellent
Q8_082.5 GB78.2 GBexcellent
F16164.4 GB156.5 GBlossless

How to run Qwen 3 Coder Next 80B-A3B on your device

Start with a short prompt and the smallest quant that still gives the quality you need. A clean first run is more useful than loading a larger file with no memory headroom.

Windows

Run Qwen 3 Coder Next 80B-A3B on Multi-GPU or High-RAM Desktop

  1. Plan for 44.7 GB minimum combined memory and 74.5 GB recommended memory for Q4_K_M.
  2. Use llama.cpp `-ot ".ffn_.*_exps.=CPU"` or `--tensor-split` across multiple GPUs so attention and KV cache remain in fast VRAM.
  3. Enable Flash Attention (`-fa on`) and Q8_0 KV cache (`-ctk q8_0 -ctv q8_0`) to prevent context buffers from exceeding GPU headroom.
  4. Avoid attempting to run workstation-scale weights on smartphones or 8 GB laptops; use a 2B–8B checkpoint on portable devices.
Device how-to guide
Mac

Run Qwen 3 Coder Next 80B-A3B on High-Memory Apple Silicon Mac

  1. Verify you are on a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra capable of holding 41.5 GB in wired unified memory.
  2. Launch via `llama-server` or LM Studio with MLX/Metal enabled and set initial context conservatively to 4K–8K.
  3. Monitor Activity Monitor Memory Pressure to confirm zero SSD swap thrashing during token generation.
  4. Step down to a lighter quantization (e.g., Q3_K_M or IQ4_XS) if total memory utilization exceeds 75%.
Device how-to guide

Qwen 3 Coder Next 80B-A3B vs alternatives

ModelParamsContextMin RAMLicense
Qwen 3 Coder Next 80B-A3B80B (3B active)256K44.7 GBApache 2.0
Qwen 3 Next 80B-A3B80B (3B active)256K44.7 GBApache 2.0
Qwen 2.5 72B72B128K40.2 GBQwen
Qwen 3.5 122B-A10B122B (10B active)256K68.2 GBApache 2.0
Qwen 3.6 35B-A3B36B (3B active)256K20.1 GBApache 2.0
Qwen 3.5 35B-A3B35B (3B active)256K19.6 GBApache 2.0

Use this table to choose the smallest nearby model that still matches the job. A related model with lower RAM can be the better local choice when the current page fits only at the edge of available memory.

Frequently asked questions

What is Qwen 3 Coder Next 80B-A3B?
Qwen 3 Coder Next 80B-A3B is a 80B sparse mixture-of-experts model (3B active active per token) from Alibaba, supporting up to 256K tokens with 44.7 GB minimum RAM and 74.5 GB recommended RAM.
How do I download and install Qwen3 coder next?
Download a GGUF build for Qwen 3 Coder Next 80B-A3B, start with Q4_K_M, and confirm the model fits locally before increasing context or quant quality.
How much VRAM or RAM does Qwen 3 Coder Next 80B-A3B require?
At Q4_K_M, Qwen 3 Coder Next 80B-A3B occupies roughly 41.5 GB of active memory and 39.1 GB on disk. Add 2–6 GB of spare headroom for KV cache buffers, runtime scratch space, and OS apps (74.5 GB recommended).
What is the best way to run Qwen 3 Coder Next 80B-A3B locally?
Use `ollama run qwen3-coder-next` for a quick CLI/API setup, or load Q4_K_M in LM Studio or llama.cpp with Flash Attention enabled when you want explicit KV cache and layer control.
Can Qwen 3 Coder Next 80B-A3B run on a consumer laptop or smartphone?
No smartphone or 8–16 GB laptop can hold Qwen 3 Coder Next 80B-A3B cleanly in local memory. It targets multi-GPU NVIDIA workstations, high-RAM Mac Studios, or MoE CPU-RAM hybrid rigs with at least 44.7 GB total memory.
How does Qwen 3 Coder Next 80B-A3B compare to Qwen 3 Next 80B-A3B?
Qwen 3 Next 80B-A3B (80B, 44.7 GB min RAM) sits closest in size. Compare active parameters, context ceiling (256K vs 256K), and license terms (Apache 2.0) before switching.