With 24 GB of VRAM, the Tesla P40 runs Qwen3 1.7B at Q4_K_M entirely on the GPU with an 8K context (about 3.1 GB needed). At Q8_0 it fits entirely.
Q4_K_M · 8K context
The model needs about 3.1 GB and fits in the 23.3 GB available, with 20.1 GB to spare.
~3.1 GB (2.8 GB–3.5 GB) needed · estimate
Fits entirely in VRAM · 20.1 GB left
EstimateData updated 24 days agoSource: Run LLM Locally engine, Hugging Face config.json + model card
What makes this estimate less reliable
| Qwen3 1.7B Q8_08K~3.7 GB · 97–137 tok/s estimate | Excellent |
|---|---|
| Qwen3 1.7B Q6_K8K~3.5 GB · 103–146 tok/s estimate | Excellent |
| Qwen3 1.7B Q5_K_M8K~3.3 GB · 112–160 tok/s estimate | Excellent |
| Qwen3 1.7B Q4_K_M8K~3.1 GB · 121–176 tok/s estimate | Excellent |
| Qwen3 1.7B IQ4_XS8K~3 GB · 128–186 tok/s estimate | Excellent |
| Qwen3 1.7B Q3_K_M8K~2.9 GB · 134–197 tok/s estimate | Excellent |
Prices and links follow the market you choose.
NVIDIA
ExcellentRuns Qwen3 1.7B Q4_K_M entirely in memory
AMD
ExcellentRuns Qwen3 1.7B Q4_K_M entirely in memory
Unified memory PC
ExcellentRuns Qwen3 1.7B Q4_K_M entirely in memory
No current offer imported for this market: the link opens a marketplace search.
