Liquid AI's d1 hits 85-97% on VisA inspection tasks with zero task-specific training
JosephJacks_ · x · 2026-10-06
Liquid AI unveiled d1, a vision-enabled decision model the company calls the best on the market. It inspects parts directly from camera images, understands tasks from a short description without task-specific training, and scores 85% to 97% across four VisA inspection tasks (circuit boards, candles, cashews, chewing gum). Observers note its decision quality rivals much larger models at roughly 200x the speed, though that claim is unverified; its small size makes local deployment plausible.
More from Models
- Attention Relay makes embedding models instruction-aware without training via LLM attention weights — _reachsumit · 2026-10-06
- OpenWork, an open-source Claude Cowork alternative, hits 45K downloads in two days — alex_verem · 2026-10-06
- Abacus.AI CEO: Chinese open-source models beat US labs on easy tasks, DeepSeek cheapest for agents — bindureddy · 2026-10-06
- HF researcher disputes GLiDE's Decision Index win over Jev as reasoning-boosted — antoine_chaffin · 2026-10-06
- ChatGPT quietly adds lifetime usage tracking to profile settings — Bpelks · 2026-10-06
- Falcon-Emirati: teaching an LLM the Emirati dialect, culture and nuance — Hugging Face Blog · 2026-10-06