← CanIRunLocal · All hardware · All models
Yes, with partial offload (slower). Qwen 2.5 32B on RX 9070 XT 16GB at Q4_K_M uses system RAM (22.2 GB needed, 15.5 GB usable), about 5 tok/s (estimate).
Hardware: RX 9070 XT 16GB · Model: Qwen 2.5 32B
RX 9070 XT 16GB has 16 GB of VRAM and 640 GB/s of memory bandwidth. The checker holds back 0.5 GB of VRAM for the driver and assumes 16 GB of system RAM, so fast memory usable by Ollama is 15.5 GB. Partial offload may also use 60% of that system RAM. Qwen 2.5 32B (32.8B, Alibaba) at Q4_K_M is a 20 GB download and needs 22.2 GB with headroom for about 8k tokens.
| Best quant | Q4_K_M |
|---|---|
| Download size | 20 GB |
| Memory needed | 22.2 GB |
| Usable fast memory | 15.5 GB |
| Fit | partial offload (slower) |
| Estimated speed | ~5 tok/s (estimate) |
| Command | ollama run qwen2.5:32b |
Estimates, not benchmarks. Model sizes are ollama.com download sizes; memory needed adds headroom for ~8k tokens of context; speeds are rough figures from memory bandwidth. Data updated 2026-10-23.
A project by OnDevice Labs · llms.txt