Can I run Qwen3.8 Flash Next locally?
Qwen3.8 Flash Next by Qwen has 180B parameters (62B active per token, 512 experts, 10 used per token). At 4-bit with an 8K context it needs about 104 GB of GPU memory. It fits on 13 of 65 GPUs and Macs we track.
Memory needed by quant and context length
GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/Qwen3.8-Flash-Next-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.
| Quant | Quality | 4K ctx | 8K ctx | 32K ctx | 128K ctx |
|---|---|---|---|---|---|
| 2-bit (Q2_K) | rough | 63.8 GB | 63.9 GB | 64.5 GB | 66.7 GB |
| 3-bit (Q3_K_M) | usable | 82.9 GB | 83.0 GB | 83.6 GB | 85.8 GB |
| 4-bit (Q4_K_M) | good, most popular | 103 GB | 104 GB | 104 GB | 106 GB |
| 5-bit (Q5_K_M) | very good | 120 GB | 121 GB | 121 GB | 123 GB |
| 6-bit (Q6_K) | near lossless | 138 GB | 139 GB | 139 GB | 141 GB |
| 8-bit (Q8_0) • | effectively lossless | 176 GB | 176 GB | 177 GB | 179 GB |
| 16-bit (F16/BF16) • | original weights | 331 GB | 331 GB | 331 GB | 334 GB |
Which GPUs and Macs run Qwen3.8 Flash Next?
Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).
Prefer an API?
Qwen: Qwen3.8 Flash is on OpenRouter at $0.15 per 1M input tokens and $0.47 per 1M output tokens. Price details and history →
Questions
How much VRAM does Qwen3.8 Flash Next need?
About 104 GB at 4-bit (Q4_K_M) with an 8K context: 102 GB of weights (estimated), 0.2 GB of KV cache and 0.9 GB of runtime overhead. At 8-bit it needs about 176 GB.
What is the smallest GPU that runs Qwen3.8 Flash Next?
No single consumer GPU in our table fits it at Q4_K_M with 8K context; the smallest option that does is the H200 (141 GB).
Can I run Qwen3.8 Flash Next on a Mac?
Yes. At Q4_K_M with 8K context it fits on a Mac M2 Ultra 192GB or larger, using the default macOS GPU memory limit.
Similar models
Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.
Qwen3.6 35B A3B
36B · 22.1 GB at 4-bitQwen3.5 35B A3B
36B · 21.5 GB at 4-bitQwen AgentWorld 35B A3B
35B · 21.6 GB at 4-bitStep 3.5 Flash
199B · 116 GB at 4-bitStep 3.7 Flash
201B · 119 GB at 4-bitMiniMax M3.1 Flash
220B · 127 GB at 4-bitModel facts
| Hugging Face repo | Qwen/Qwen3.8-Flash-Next |
| Parameters | 179,999,981,459 (180B) · ~62B active (computed from config) |
| Layers / hidden size | 48 / 2560 |
| Attention | GQA (hybrid: 36 linear/state layers); 24 heads, 2 KV heads, head dim 256 |
| KV cache per token (FP16) | 24.0 KiB |
| Max context | 262,144 tokens |
| License | other |
| Released on HF | Aug 24, 2026 |
| Downloads (30d) / likes | 2,975,526 / 7,230 |
Sources: Hugging Face model API · config.json (Qwen/Qwen3.8-Flash-Next) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.