Is d1-2B state of the art?Pending benchmarks.Announced Oct 27, 2026; no independent results yet.
Liquid AI's decision model for text, pictures and video, built on LFM3-VL-2B. Instead of generating text, it answers any number of named questions about one input in a single forward pass, with typed answers (yes/no, choices, scores, extracted values, boxes, estimates) and calibrated probabilities. Open weights under the LFM Open License v1.0.
The details so far
- Lab
- Liquid AI
- Announced
- Oct 27, 2026
- API
- Live · Liquid AI API and Playground, per Liquid's blog
- Open weights
- Published
- Parameters
- 1.92B
- Base model
- LFM3-VL-2B
- Input
- A state (text or JSON), pictures or a video, and any number of named questions
- Output
- Typed answers with probabilities; no generated tokens. Not a chat model
- Context
- 128K tokens (Liquid says answers get less reliable past about 64K tokens of state)
- License
- LFM Open License v1.0
Tags marked unsure are not stated on the lab's own pages.
What would let us call it
Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.
- Publicly announced with a source
- Usable by the public (API or weights)
- Listed on an independent board we track
- Enough results for a rank in a category
- Holds across two snapshots (crowns only)
What the lab claims
Self-reported. Shown for context; not used in any verdict.
Lab claim: replaces d1-3B with a third fewer parameters, scoring 85.2 vs 83.6 on Decision Bench and 75.5 vs 64.0 on ImajevBench; 5.7 ms per decision on an NVIDIA H100.
Sources
LABLiquid AI: Introducing LFM3liquid.ai/blog/lfm3 · Oct 27, 2026REPOLiquidAI/d1-2B on Hugging Facehuggingface.co/LiquidAI/d1-2BTracking continues until an independent board lists it.