← Latest products

Cactus Hybrid

We taught Gemma 4 to know when it's wrong. On-device models that know when they're wrong: every answer carries a confidence score for cloud handoff. A small, on-device model is fast and private, but sometimes wrong, but frontier models are getting expensive pretty fast.

Open source

271 stars · GitHubexternal site

Visit github.comexternal site

Reactions (1)

Similar products

by description similarity

  1. Cactus (YC S25)

    github.com

    AI inference on smartphones

    Hacker News231 pts

  2. ChonkLM

    chonklm.com

    Tiny language models running offline in the browser

    Hacker News6 pts

  3. General Instinct (YC P26)

    general-instinct.com

    Frontier models on edge devices

    Hacker News63 pts

  4. Bonsai 1.7B ternary model at 442T/s on M4 Max

    agents2agents.ai

    A drop-in optimized inference build of Bonsai 1.7B for Apple Silicon. ~42% faster decode (tg128) and ~9% faster prefill (pp512) on M-series GPUs. All Metal kernels written and tuned by ata, our autonomous engineering agent.

    Hacker News13 pts

  5. I benchmarked Gemma 4 E2B

    aiexplr.com

    the 2B model beat the 12B on multi-turn

    Hacker News8 pts