What GPU do you need to run Qwen3.8 Flash Next?
180B params 256K max context VisionTool useReasoning
The smallest GPU memory that runs Qwen3.8 Flash Next well is 128 GB (Q4_K_M, e.g. RTX Spark). 11 of 132 GPUs run it well. It's a mixture-of-experts model, so all weights must fit in memory but speed depends only on the active experts.
MMLU 73.23
VRAM needed per quantization
| Quantization | Weights | 16K context | 32K context | Full 256K |
|---|---|---|---|---|
| Q4_K_M | 111.4 GB | 111.8 GB | 112.1 GB | 117.3 GB |
| Q8_0 | 175.3 GB | 175.7 GB | 176.1 GB | 181.2 GB |
Weights plus fp16 KV cache. Leave ~1 GB headroom for the inference engine.
Cheapest GPUs that can run Qwen3.8 Flash Next
Lowest launch MSRP among GPUs that run its smallest quantization (Q4_K_M) well, with at least 16K context.
As an Amazon Associate, I earn from qualifying purchases.
GPUs that can run Qwen3.8 Flash Next
Best quantization that fits on each GPU, with the context and speed it gets. For chips with several memory sizes, the smallest that works is shown.
NVIDIA
- DGX Spark (GB 10) 128 GB Q4_K_M 256K context · ~58 tok/s
- RTX Spark 128 GB Q4_K_M 256K context · ~64 tok/s
AMD
- Ryzen AI Max (Strix Halo) 128 GB Q4_K_M 256K context · ~54 tok/s
- Ryzen AI Max 400 (Gorgon Halo) 128 GB Q4_K_M 256K context · ~58 tok/s
Apple
- M1 Ultra 128 GB Q4_K_M 256K context · ~170 tok/s
- M2 Ultra 128 GB Q4_K_M 256K context · ~170 tok/s
- M3 Max 128 GB Q4_K_M 256K context · ~85 tok/s
- M4 Max 128 GB Q4_K_M 256K context · ~116 tok/s
- M5 Max 128 GB Q4_K_M 256K context · ~130 tok/s
- M3 Ultra 256 GB Q8_0 256K context · ~120 tok/s
- M5 Ultra 256 GB Q8_0 256K context · ~175 tok/s