Models

Popular models from Hugging Face that GridPup can serve, grouped by family and size, with the GPU memory each one needs and the cheapest machine that fits it right now. Open a family to compare its sizes and versions (the original, quantized and ONNX versions need less GPU memory) and the GPUs for each.

1,050 families · catalog updated 12:56 UTC

• Phi-3-medium-4k by microsoftText generation · 1 model · ↓ 19.2K~14B
  Smallest needs34.2 GB GPU
  From$0.41/h4x RTX 3060
• Olmo-3.1 by allenaiText generation · 1 model · ↓ 19K32B
  Smallest needs76.2 GB GPU
  From$0.89/hNvidia A100
• Yi-Coder by 01-aiText generation · 1 model, 2 versions · ↓ 18.8K9B
  Smallest needs22.7 GB GPU
  From$0.23/hRTX 3090
• Qwen2.5 by mazdaypciText generation · 1 model · ↓ 18.5K0.5B
  Smallest needs3.4 GB GPU
  From$0.09/hRTX 3060
• gemma-4 by DavidAUText generation · 1 model, 2 versions · ↓ 18.4K31B
  Smallest needs33.9 GB GPU
  From$0.41/h4x RTX 3060

Model details come from Hugging Face and are refreshed daily. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090"). Only machines with a Vast reliability above 95% that download at 400 Mbit/s or more are counted.