DeepSeek-V4-Pro (NVFP4-FP8 (compressed-tensors))

DeepSeek · DeepSeek
LLMMITpublicTrusted1600B params

Ready-made NVFP4-FP8 (compressed-tensors) build - runs on the fleet stack (compressed-tensors). Blackwell window.

Ready-made NVFP4-FP8 (compressed-tensors) build - runs on the fleet stack (compressed-tensors). Blackwell window.

At a glance

  • Task: LLM (InferenceV1)
  • Params: 1600B
  • License: MIT (commercial use OK)
  • Hardware: 1003 GB+ VRAM
  • dtype: Nvfp4