LLM Torrents
Models Collections Learn Hardware Log in

Collections

Curated, editorial lists of models — by size, use case, and hardware.

  • Best under 8 GB

    2 models

    Capable models that comfortably fit an 8 GB memory budget at a 4-bit quant — a good starting point for laptops and entry GPUs.

  • Best coding (≤ 24 GB)

    0 models

    Coding-focused models sized to run on a 24 GB card at a 4-bit quant. Selection is editorial; see each model page for license and provenance.

  • Best on Apple Silicon

    2 models

    Models that run well on unified-memory Apple Silicon Macs. Larger sizes assume more unified memory; check the VRAM guide for how memory is s...