Build: NVIDIA GB200 NVL72 + — + 32 GB RAM (windows)
Ranked by fit for vision use case + predicted speed. Click a row for VRAM breakdown.
ollama run gemma4:e2bollama run gemma4:e4bollama run gemma3:4bollama run llama3.2-vision:11bollama run gemma3:12bollama run pixtral:12bollama run gemma4:26b-moeollama run gemma3:27bollama run gemma4:31bHypothetical scenarios. We re-ran the compatibility engine for each.
~$80–150
Doubles your CPU-offload working set. Helps when models don't quite fit in VRAM.
Unlocks: 149 new comfortable
see current pricing
Tensor parallelism splits the model across both cards, effectively doubling VRAM. Bandwidth doesn't double — runs ~1.5× the single-card speed in practice.
Unlocks: 149 new comfortable
Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.
Need more memory than you have. Shown for orientation.
Even with CPU offload, needs more memory than your VRAM (13824 GB) + 60% of system RAM (19 GB) combined.
Even with CPU offload, needs more memory than your VRAM (13824 GB) + 60% of system RAM (19 GB) combined.
Even with CPU offload, needs more memory than your VRAM (13824 GB) + 60% of system RAM (19 GB) combined.
Want a specific benchmark we don't have? Email support@runlocalai.co and we'll prioritize it.