LFM-3B: benchmark results
Liquid AI's original September 2024 LFM-3B, released through hosted and commercial access without public weights. Later open LFM generations are separate releases. Provider: Liquid AI. Released 2024-09-30. Access: API.
Aggregate Rating 1365 ± 37, rank #1351 of 1656 rated models, from 40 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AgentDrive - Scenario | 92.5 | Accuracy (%) | 50 |
| AgentDrive - Physics | 40 | Accuracy (%) | 49 |
| AgentDrive - Hybrid | 25 | Accuracy (%) | 39 |
| MERA - BPS | 89.7 | Accuracy (%) | 36.7 |
| MERA - PARus | 71.2 | Accuracy (%) | 34.6 |
| MERA - ruWorldTree | 78.29 | Accuracy (%) | 34.3 |
| MERA - ruTiE | 61.91 | Accuracy (%) | 33.8 |
| MERA - ruOpenBookQA | 68.75 | Accuracy (%) | 33.3 |
| MERA - RCB | 50.23 | Accuracy (%) | 31 |
| MERA - USE | 10 | Grade, normalized (%) | 30.5 |
| MERA - MultiQ | 26.55 | F1 (%) | 29.4 |
| MERA - ruMMLU | 45.51 | Accuracy (%) | 28.1 |
Interactive version: theaggregate.ai/model?slug=lfm-3b · How It Works · Data refreshed daily, snapshot 2026-10-20.