Alibaba (Qwen) · Qwen Community
Qwen3.8 Flash Next
An experimental preview of the next Qwen design, and the most-used model on 128 GB machines. Smaller 3-bit and 4-bit files fit 96 GB.
How much memory it needs
The model file, plus memory for the conversation, plus about 1 GB for the app. Longer conversations need more.
| Size | Version | File | At 8K | At 32K | At 128K |
|---|---|---|---|---|---|
| 180B (6B active) | UD-IQ3_XXS | 82.0 GB | 84.1 GB | 84.7 GB | 87.1 GB |
| UD-IQ4_XS | 93.7 GB | 95.8 GB | 96.4 GB | 98.8 GB | |
| Q4_K_M | 120 GB | 122 GB | 122 GB | 125 GB | |
| Q8_0 | 188 GB | 190 GB | 191 GB | 193 GB |
Which computers run it
The best size for each machine at an 8K-token conversation. Speeds are estimates from memory bandwidth.
| Computer | Best version | Fit | Speed |
|---|---|---|---|
| Laptop or PC, no graphics card, 32 GB | 180B UD-IQ3_XXS | Too big | — |
| RTX 4060 · 8 GB | 180B UD-IQ3_XXS | Too big | — |
| RTX 3060 · 12 GB | 180B UD-IQ3_XXS | Too big | — |
| RTX 5060 Ti · 16 GB | 180B UD-IQ3_XXS | Too big | — |
| RTX 4090 · 24 GB | 180B UD-IQ3_XXS | Too big | — |
| RTX 5090 · 32 GB | 180B UD-IQ3_XXS | Too big | — |
| Mac with M4 · 16 GB | 180B UD-IQ3_XXS | Too big | — |
| Mac with M4 Pro · 48 GB | 180B UD-IQ3_XXS | Too big | — |
| Mac with M5 Max · 128 GB | 180B UD-IQ4_XS | Runs well | 50–110tok/sFast |
| Ryzen AI Max+ 395 · 128 GB | 180B UD-IQ3_XXS | Runs well | 30–60tok/sFast |
| DGX Spark · 128 GB | 180B UD-IQ4_XS | Runs well | 30–55tok/sFast |
| RTX PRO 6000 · 96 GB | 180B UD-IQ3_XXS | Runs well | 120–240tok/sFast |
| Mac Studio with M5 Ultra · 256 GB | 180B Q8_0 | Runs well | 35–70tok/sFast |
| Mac Studio with M3 Ultra · 512 GB | 180B Q8_0 | Runs great | 25–55tok/sFast |
What it is good and bad at
Good at
- Fast for its size: only 6B parameters work per word
- Very small conversation memory
Not good at
- Even the 3-bit file is 82 GB, so it needs 96 GB or more
- Licence limits some large commercial uses
Start it
Install Ollama, then run one of these in a terminal. LM Studio users can search the same name in the app.
ollama run qwen3.8-flash-next:125b-a6b-q4_K_M