Qwen2.5 VRAM requirements
Memory needed to run Qwen2.5 at every published size and quant, with an 8k context. Each figure is weights plus KV cache plus runtime buffers, from the same memory model as the VRAM calculator.
Total memory by size and quant
GB at an 8k context. Bold is the Q4_K_M column, the usual starting point.
| Size | Q8_0 | Q6_K | Q5_K_M | Q4_K_M | Q3_K_M |
|---|---|---|---|---|---|
| 7B | 10.0 | 8.7 | 7.8 | 6.9 | 6.0 |
| 27B | 31 | 26 | 22 | 19 | 16 |
| 32B | 36 | 30 | 26 | 22 | 18 |
KV cache by context length
GB the context alone adds on top of the weights. Add this to a weights figure to size a longer session.
| Size | Weights at Q4_K_M | 4k ctx | 16k ctx | 32k ctx | 128k ctx |
|---|---|---|---|---|---|
| 7B | 3.9 | 0.6 | 2.4 | 4.8 | 19 |
| 27B | 15 | 1.1 | 4.3 | 8.6 | 34 |
| 32B | 18 | 1.1 | 4.3 | 8.6 | 34 |
What each size fits on at Q4_K_M
Hardware whose memory holds the model with headroom, smallest pool first. Fits means under 85% of the pool at an 8k context.
- 7Bneeds about 6.9 GB
Smallest: NVIDIA P102-100 (10 GB), RTX 2080 Ti (11 GB), D700 12GB (12 GB), Intel Arc B580 12GB (12 GB), and 93 more.
- 27Bneeds about 19 GB
Smallest: L4 (24 GB), M4 24GB (24 GB), M4 Pro 24GB (24 GB), RTX 3090 (24 GB), and 60 more.
- 32Bneeds about 22 GB
Smallest: AMD MI50 32GB (32 GB), M1 Max 32GB (32 GB), M1 Pro 32GB (32 GB), M2 Max 32GB (32 GB), and 53 more.
Frequently asked
How much VRAM does Qwen2.5 need?
From about 6.9 GB for the 7B model to about 22 GB for the 32B model, at Q4_K_M with an 8k context and including the KV cache and runtime buffers. Q8 needs more and Q3 less; the table above lists every size and rung.
Can I run Qwen2.5 on a 16GB GPU?
Yes. The 7B model needs about 6.9 GB at Q4_K_M with an 8k context, which fits a 16 GB card with headroom. Larger sizes need a smaller quant or a bigger card.
Can I run Qwen2.5 on a 24GB GPU such as an RTX 3090 or 4090?
Yes. The 27B model needs about 19 GB at Q4_K_M with an 8k context, which fits a 24 GB card with headroom. Larger sizes need a smaller quant or more memory.
How are these numbers calculated?
Weights at the quant's bits per weight, plus a KV cache sized from the model's attention design for the chosen context, plus compute buffers and a fixed runtime allowance. It is the same memory model the VRAM calculator uses. A family whose attention profile is not on file is sized as plain grouped-query attention, which errs towards needing more memory.