Qwen2.5
model paperThe "18-trillion token" series. Released sizes: 0.5B, 1.5B, 3B, 7B, 14B, 32B, and 72B, plus a proprietary MoE model (Qwen2.5-Max).
Outputs 3
Qwen2.5
model Architecture DENSE
Variants
| Name | Parameters | Notes |
|---|---|---|
| Qwen2.5-0.5B | 0.5B | — |
| Qwen2.5-1.5B | 1.5B | — |
| Qwen2.5-3B | 3B | — |
| Qwen2.5-7B | 7B | — |
| Qwen2.5-14B | 14B | — |
| Qwen2.5-32B | 32B | — |
| Qwen2.5-72B | 72B | — |
Qwen2.5 Technical Report
paperFull Qwen2.5 series technical report. Pre-trained on 18T tokens.
arXiv: 2412.15115
Qwen2.5-Max
modelProprietary-grade MoE model served via API, competing with GPT-4o.
Architecture MOE