Our Selection

The models we recommend running locally, hand-picked across the labs.

Start with one of these

  • Qwen3 · 0.6B

    Alibaba Cloud

    Text generation

    Released in September 2025

    0.6B 0.48 GB Smartphone Thinking 110 languages
  • Qwen3 · 4B

    Alibaba Cloud

    Text generation

    Released in September 2025

    4B 2.5 GB Laptop Thinking 110 languages

All models