Explore AI
48 GB VRAM or more
Large open models, no compromises.
Large open models, no compromises.
- Platform
- NVIDIA
- Memory
- 48 GB vram
Multi-GPU setups or professional cards. At this point compare the electricity bill against renting a GPU by the hour — it is often cheaper to rent.
Runs comfortably
- GPT-OSS 120B 64 GB at Q4 · 120B total, 5.1B active Apache-2.0
- Qwen3 32B 24 GB at Q4 Apache-2.0
- Qwen3 Coder 30B A3B 20 GB at Q4 · 30B total, 3B active Apache-2.0
Worth knowing
Renting a cloud GPU for a few hours a month usually beats buying hardware at this tier. Do the arithmetic before committing.