← CanIRunLocal · All hardware · All models
Yes, with partial offload (slower). Qwen3.8 27B on RTX 5080 16GB at Default uses system RAM (20.3 GB needed, 15.5 GB usable), about 8 tok/s (estimate).
Hardware: RTX 5080 16GB · Model: Qwen3.8 27B
RTX 5080 16GB has 16 GB of VRAM and 960 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. Qwen3.8 27B (27B, Alibaba) at Default is a 18 GB download and needs 20.3 GB with headroom for about 8k tokens.
| Best quant | Default |
|---|---|
| Download size | 18 GB |
| Memory needed | 20.3 GB |
| Usable fast memory | 15.5 GB |
| Fit | partial offload (slower) |
| Estimated speed | ~8 tok/s (estimate) |
| Command | ollama run qwen3.8:27b |
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