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

• gpt2-small-dutch by GroNLPText generation · 1 model · ↓ 17.4K~129M
  Smallest needs2.5 GB GPU
  From$0.10/hRTX 3060
• Llama-3-Instruct by grimjimText generation · 2 models · ↓ 17.3K8B
  Smallest needs21.3 GB GPU
  From$0.21/h2x RTX A4000
• Wald by org2aiText generation · 1 model · ↓ 17.3K4B
  Smallest needs13.8 GB GPU
  From$0.15/hRTX 5060 Ti
• Llama-3-Instruct by haoranxuText generation · 2 models · ↓ 17.3K8B
  Smallest needs21.3 GB GPU
  From$0.21/h2x RTX A4000
• ChimeraLlama-3 by mlabonneText generation · 2 models · ↓ 17.3K8B
  Smallest needs21.3 GB GPU
  From$0.21/h

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.