No — not on this device
Qwen3.8-Flash-Next 180B-A6B at Q4_K_M needs 102.0 GB against 5.4 GB usable, and the shortfall of 96.6 GB is more than 32 GB of system RAM can cover at a tolerable speed. A smaller sibling or a lower quantisation is the honest answer here.
The open preview of the Qwen4 architecture: a 125B-A6B hybrid (Gated DeltaNet + sparse attention, KV cache on 12 of 48 blocks) plus a 51B n-gram embedding and a 4B draft head — 180B on disk, 6B active. The hosted "Qwen3.8-Flash" is this model with a 1M window.
Hardware requirements What hardware do I need to run Qwen3.8-Flash-Next 180B-A6B? →At Q4_K_M with an 8K context: a 141 GB GPU (H200 SXM) or a 192 GB Mac (M2 Ultra). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.
The VRAM budget
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 178.1 GB | 178.9 GB | — | ~3.0 | −0.1% ppl | 173.5 GB over |
| Q6_K | 137.5 GB | 138.3 GB | — | ~3.9 | −0.4% ppl | 132.9 GB over |
| Q5_K_M | 118.8 GB | 119.6 GB | — | ~4.5 | −0.8% ppl | 114.2 GB over |
| Q4_K_M | 101.2 GB | 102.0 GB | — | ~5.3 | −1.9% ppl | 96.6 GB over |
| Q3_K_M | 81.9 GB | 82.7 GB | — | ~6.5 | −5.4% ppl | 77.3 GB over |
| Q2_K | 70.2 GB | 71.0 GB | — | ~7.6 | −15% ppl | 65.6 GB over |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. Only 12 of its 48 blocks keep a per-token KV cache; the rest are linear-attention, Mamba or convolution blocks with a fixed-size state.
How to run it
$ pip install mlx-lm $ mlx_lm.generate --model mlx-community/Qwen3.8-Flash-Next-4bit \ --max-tokens 512 --prompt "Hello"
Apple's own array framework. The fastest path on Apple Silicon. More on MLX.
Related models on an M1 · 8 GB
| Model | Total | Tok/s est. | |
|---|---|---|---|
| Qwen3.8 27B 27.8B · same family |
Q4_K_M · 16.7 GB | — | Won't fit |
| Qwen3.8 2.4T-A95B 2446B MoE · same family |
Q4_K_M · 1376.7 GB | — | Won't fit |
| Qwen3 235B-A22B 235B MoE · similar size |
Q4_K_M · 134.2 GB | — | Won't fit |
| Qwen3.5 122B-A10B 125B MoE · similar size |
Q4_K_M · 71.1 GB | — | Won't fit |
| Nemotron 3 Super 120B-A12B 124B MoE · similar size |
Q4_K_M · 70.4 GB | — | Won't fit |
| Ling 3.0 Flash 124B-A5B 124B MoE · similar size |
Q4_K_M · 70.4 GB | — | Won't fit |
| Mistral Small 4 119B-A6B 119B MoE · similar size |
Q4_K_M · 67.7 GB | — | Won't fit |
Qwen3.8-Flash-Next 180B-A6B on other hardware
| Device | Total | Tok/s est. | |
|---|---|---|---|
| GeForce RTX 4060 Laptop 7.0 GB usable · 256 GB/s |
102.0 GB | — | Won't fit |
| GeForce RTX 3060 12 GB 10.6 GB usable · 360 GB/s |
102.0 GB | — | Won't fit |
| M4 · 16 GB 10.7 GB usable · 120 GB/s |
102.0 GB | — | Won't fit |
| GeForce RTX 4090 22.4 GB usable · 1008 GB/s |
102.0 GB | — | Won't fit |
| Radeon RX 7900 XTX 22.4 GB usable · 960 GB/s |
102.0 GB | — | Won't fit |
| M3 Max · 64 GB 48.0 GB usable · 400 GB/s |
102.0 GB | — | Won't fit |
| H200 SXM 139.4 GB usable · 4800 GB/s |
102.0 GB | 331 | Runs great |
| M2 Ultra · 192 GB 144.0 GB usable · 800 GB/s |
102.0 GB | 62 | Runs great |