DeepSeek-V4.1-Flash-Ternary

  • - for complex coding and research agents
  • - ~92% intelligence retained vs FP4
  • - 141 GB pack · 28% of 510 GB base
  • - ~12.1x speedup

Qwen3.8-Flash-Next-NVFP4

  • - for coding and everyday use
  • - ~99.9% intelligence retained vs BF16
  • - 187 GB · 52% of 360 GB BF16
  • - ~7.8x speedup

Qwen3.8-27B-NVFP4-NInfer

  • - for personal agents
  • - ~99% intelligence retained vs BF16
  • - 24 GB · 43% of 56 GB BF16
  • - ~4.5x speedup

Qwen3.8-Flash-Next-Ternary

  • - for fast retrieval and data analysis
  • - ~91% intelligence retained vs BF16
  • - 32 GB pack · 9% of 360 GB BF16
  • - ~1.85x speedup