Qwen3-Coder-Next (80B-A3B MoE) — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
Qwen's efficiency-first coder, released as Qwen3-Coder-Next: 80B total with 3B active parameters per token, built on the hybrid-attention Qwen3-Next-80B-A3B base (most layers keep no KV cache, so long context is cheap). All 80B parameters must sit in memory, so it needs a 64 GB-class machine at Q4; only the 3B active experts are read per token, which keeps decoding fast. Apache 2.0, 256K context. Published SWE-bench figures disagree between sources, so none is quoted here.
Qwen3-Coder-Next (80B-A3B MoE) needs about 49 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
VRAM and speed by quantization
Quoted against NVIDIA RTX 4090 (24 GB). Includes the KV cache at 8K context, so it reads higher than the headline figure.
| Quant | Memory | VRAM | Speed (est.) | Fit |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 27.4 GB | ~40 tok/s | Offload | |
| Q3_K_M 3.41 bpw | 35.2 GB | ~36 tok/s | Offload | |
| Q4_K_M 4.83 bpw | 49.4 GB | ~31 tok/s | Offload | |
| Q5_K_M 5.67 bpw | 57.8 GB | — | Too big | |
| Q6_K 6.56 bpw | 66.7 GB | — | Too big | |
| Q8_0 8.50 bpw | 86.1 GB | — | Too big | |
| F16 16.00 bpw | 161.1 GB | — | Too big |
Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Qwen3-Coder-Next (80B-A3B MoE) VRAM calculator →
Get Qwen3-Coder-Next (80B-A3B MoE) running
The cheapest catalogued GPU that runs Qwen3-Coder-Next (80B-A3B MoE) is the NVIDIA DGX Spark (128 GB).
or compare on Vast.ai from $0.77/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
How to run Qwen3-Coder-Next (80B-A3B MoE)
Install Ollama, then run:
ollama run qwen3-coder:80b-a3b-q4Specifications
Curated — A hand-written entry from before this catalogue recorded its sources. The figures are long-standing but their provenance is not on file.
- Parameters
- 80 Billion (3B active)
- Context window
- 262,144
- Architecture
- MoE (hybrid attention)
- Provider
- Alibaba Cloud
- Licence
- Apache 2.0
- Practical deployment
- Workstation (large unified memory or pro GPU)
- Specified at
- Q4_K_M
- System RAM
- 32 GB
- Record updated
- 2026-09-29
Commercial use permitted. No usage restrictions beyond attribution.
Quality and use cases
Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.
Can I run Qwen3-Coder-Next (80B-A3B MoE) on my GPU?
- Qwen3-Coder-Next (80B-A3B MoE) on AMD Radeon RX 9060 XT 8GB
- Qwen3-Coder-Next (80B-A3B MoE) on AMD Ryzen AI Max+ 395
- Qwen3-Coder-Next (80B-A3B MoE) on Apple M2 Max
- Qwen3-Coder-Next (80B-A3B MoE) on Apple M3 Ultra
- Qwen3-Coder-Next (80B-A3B MoE) on Intel Arc B570
- Qwen3-Coder-Next (80B-A3B MoE) on Intel Arc B580
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA A100 80GB (PCIe)
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 3060 (12GB)
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 3070
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 3070 Ti
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 3080 (10GB)
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 4060
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 4070
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 4070 Super
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 4070 Ti
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 5060
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 5060 Ti 8GB
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA GeForce RTX 5070
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA H100 80GB (PCIe)
- Qwen3-Coder-Next (80B-A3B MoE) on NVIDIA RTX PRO 6000 Blackwell
Other Qwen3-Coder sizes
Qwen3-Coder-Next (80B-A3B MoE) — frequently asked questions
How much VRAM does Qwen3-Coder-Next (80B-A3B MoE) need?
About 49 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does Qwen3-Coder-Next (80B-A3B MoE) run on an RTX 4090 (24 GB)?
No. Qwen3-Coder-Next (80B-A3B MoE) needs about 49 GB at Q4_K_M, more than a single RTX 4090's 24 GB. It needs a larger card, several GPUs, or Apple Silicon with enough unified memory — or it runs with part of the weights offloaded to system RAM, which is much slower.
How do I run Qwen3-Coder-Next (80B-A3B MoE) locally?
Install Ollama and run `ollama run qwen3-coder:80b-a3b-q4`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Qwen3-Coder come in?
Qwen3-Coder 8B (6 GB), Qwen3-Coder 30B-A3B (MoE) (19 GB), Qwen3-Coder-Next (80B-A3B MoE) (49 GB), Qwen3-Coder 480B-A35B (MoE) (291 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.