Free {family} model via API
Liquid AI's LFM models are compact non-transformer foundation models designed for on-device speed.
Read more
The table refreshes every four hours from aggregator APIs. The same model at different providers is a separate endpoint with its own limits.
| # | Model ↕ | AA Index ↕ | GPQA ↕ | Context ↕ | Size ↕ | Free at ↕ | |
|---|---|---|---|---|---|---|---|
| 1 | 8.4 | 56% | 128K | Small |   OpenRouter, UnoRouter |
||
| 2 |
Lfm2.5 24B
Liquid AI · 24B · 2B · New
|
— | — | 128K | Mid-size | UnoRouter |
|
| 3 |
LFM3-1.7B
Liquid AI · 1.7B · New
|
— | — | 262K | Small | OpenRouter |
Benchmarks: Artificial Analysis (September 24, 2026). A dash means the model is not on their leaderboard; we never estimate scores ourselves.
About LFM
Liquid AI's LFM models are compact non-transformer foundation models designed for on-device speed.
FAQ
Which LFM models are free?
Currently 3 models are free: LFM2.5-2.6B, Lfm2.5 24B, LFM3-1.7B.
Where can I get a free LFM API key?
Free LFM routes are available at: OpenRouter, UnoRouter. Get a key in the provider's dashboard; links are on each model card.
Who develops LFM?
Liquid AI's LFM models are compact non-transformer foundation models designed for on-device speed.