Qwen3-Next 80B-A3B (Thinking)
Qwen3-Next 80B-A3B (Thinking) is an open-weights language model from Qwen, released September 1, 2025. It has 80B parameters (mixture-of-experts, 3B active). The context window is 128K tokens. API price: $0.5 input / $6 output per million tokens.
We believe that Context Length Scaling and Total Parameter Scaling are two major trends in the future of large models. To further improve training and inference efficiency under long-context and large-parameter settings, we design a brand-new model architecture called Qwen3-Next.Compared to the MoE structure of Qwen3, …
Check it against your own PC →
Can you run Qwen3-Next 80B-A3B (Thinking) locally?
The smallest common GPU it runs on is 6 GB of VRAM (with 32 GB of RAM): IQ2_M, GPU + experts in RAM, roughly 40 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Runs · GPU + CPU experts | IQ2_M (heavy loss) | 5.2 GB | 25.0 GB | ~40 |
| 8 GB VRAM | Runs · GPU + CPU experts | IQ2_M (heavy loss) | 7.2 GB | 23.0 GB | ~45 |
| 12 GB VRAM | Runs · GPU + CPU experts | IQ3_XXS (noticeable loss) | 11.2 GB | 22.6 GB | ~46 |
| 16 GB VRAM | Runs · GPU + CPU experts | IQ3_XXS (noticeable loss) | 15.2 GB | 18.6 GB | ~54 |
| 24 GB VRAM | Runs · GPU + CPU experts | IQ4_XS (good) | 23.2 GB | 22.5 GB | ~48 |
| 32 GB VRAM | Runs · GPU + CPU experts | Q4_K_M (very good) | 31.2 GB | 20.9 GB | ~53 |
Benchmark results
Epoch AI has not published results for Qwen3-Next 80B-A3B (Thinking) yet. New models are usually evaluated within days to weeks.
Links
More from Qwen
Qwen3.8 Omni Flash · Qwen3.8 Max (0902) · Qwen3.8 Flash · Qwen3.8 Flash Next · Qwen 3.8 27B · Qwen3.8-2.4T-A95B · Qwen 3.8 Max · Qwen3.7 Flash