LFM3-VL-2B: benchmark results
Provider: Liquid AI. Released 2026-10-27. Access: Open.
Aggregate Rating 1569 ± 34, rank #1418 of 10916 rated models, from 97 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LFM3-VL-2B Release - ActivityNet-QA (test) | 71.2 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - BLINK | 71.1 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - CV-Bench | 88.8 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - ChartQA (test) | 92.7 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - EmbSpatial-Bench | 82.9 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - MindCube (tinybench) | 56.5 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - Point-Bench | 73.8 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - RefCOCO-M | 95.2 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - RefCOCO-avg | 90.5 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - SAT (val + test) | 80.2 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - ScreenSpot-v2 Desktop | 95.5 | Published normalized score (0–100) | 100 |
| LFM3-VL-2B Release - VSI-Bench | 63.3 | Published normalized score (0–100) | 100 |
Interactive version: theaggregate.ai/model?slug=lfm3-vl-2b · How It Works · Data refreshed daily, snapshot 2026-11-08.