Show HN: Vram.run – Compare API providers, local GPUs, and cloud for any model

2 comments

• jad-nohra 5 hours ago

  [dead]

  [-]
  • ranger_danger 3 hours ago

    The data for Llama-4-Scout seem quite wrong to me... it says a 17B model is only 7B, and that Q4 is 4GB when it's more like 60GB.