Olmo 3.1 32B Instruct at Q4_K_M needs about 21.2 GB, more than the 16 GB of the RX 9070 XT: part of the model runs from system RAM, which slows generation. At Q3_K_M it does not fit in VRAM.
Q4_K_M · 8K context
The model needs about 21.2 GB but only 15.5 GB of VRAM is usable: 5.7 GB must be loaded into system RAM. It runs, 73% on the GPU, but more slowly.
~21.2 GB (20.5 GB–21.9 GB) needed · estimate
5.7 GB must be offloaded to system RAM
EstimateData updated 40 days agoSource: Run LLM Locally engine, Hugging Face config.json + model card
What makes this estimate less reliable
| Olmo 3.1 32B Instruct Q8_08K~35.6 GB · 1–1.9 tok/s estimate | Not recommended |
|---|---|
| Olmo 3.1 32B Instruct Q6_K8K~28 GB · 1.5–3 tok/s estimate | Possible but slow |
| Olmo 3.1 32B Instruct Q5_K_M8K~24.5 GB · 2–4 tok/s estimate | Possible but slow |
| Olmo 3.1 32B Instruct Q4_K_M8K~21.2 GB · 3–6 tok/s estimate | CPU/GPU offload |
| Olmo 3.1 32B Instruct IQ4_XS8K~19.1 GB · 4.5–9 tok/s estimate | CPU/GPU offload |
| Olmo 3.1 32B Instruct Q3_K_M8K~17.4 GB · 7.5–14 tok/s estimate | CPU/GPU offload |
Prices and links follow the market you choose.
AMD
CompatibleRuns Olmo 3.1 32B Instruct Q4_K_M entirely in memory
Apple Silicon
CompatibleRuns Olmo 3.1 32B Instruct Q4_K_M entirely in memory
Unified memory PC
CompatibleRuns Olmo 3.1 32B Instruct Q4_K_M entirely in memory
No current offer imported for this market: the link opens a marketplace search.
