LLM AtlasLLM AtlasSearch models

Provider analysis

Liquid AI

Publisher of LFM2 hybrid models designed for on-device deployment. MoE architectures from 350M to 24B parameters with edge-optimized inference.

Last verified: 2026-03-29Confidence: HighPrimary sources: 3

This provider page blends full-profile entries with broader verified listings. Use it to separate deeply evaluated flagship models from source-backed records that are tracked primarily for market visibility, access data, and freshness coverage.

Headquarters
Boston, MA
Founded
2023
Models tracked
5
Full-profile models
5
Catalog last verified
2026-03-29
Latest model verification
2026-03-29
Newest release tracked
2026-02-24
Confidence
High
Access mix
open-weight, self-hosted
API models
0

Tracked models available through provider-managed APIs.

Open-weight models
5

Models with downloadable weights or self-hosted distribution paths.

Primary source links
20

Total source references attached across this provider catalog.

Provider sources

Official links used to verify the provider profile and platform coverage.

Last verified: 2026-03-29

Liquid AI

LFM2-24B-A2B

LFM2

Liquid AI's flagship 24B/2.3B-active MoE hybrid model. 112 tok/s on AMD CPU, 293 tok/s on H100. Fits in 32GB RAM. Trained on 17T tokens with 30 conv + 10 attn layers. Supports 9 languages.

Score 604 sources
texttool-useopen-sourceopen-weightself-hosted
Context
32,768
Input
Not applicable
Output
Not applicable
Coverage
Full profile

Liquid AI

LFM2-8B-A1B

LFM2

Liquid AI's 8.3B/1.5B-active MoE hybrid model with 12T training tokens. 24 layers (18 conv + 6 attn). 47.9K downloads on HuggingFace.

Score 554 sources
texttool-useopen-sourceopen-weightself-hosted
Context
32,768
Input
Not applicable
Output
Not applicable
Coverage
Full profile

Liquid AI

LFM2-2.6B

LFM2

Liquid AI's 3B parameter LFM2 for lightweight text generation and tool use on edge devices.

Score 474 sources
textopen-sourceopen-weightself-hosted
Context
32,768
Input
Not applicable
Output
Not applicable
Coverage
Full profile

Liquid AI

LFM2.5-1.2B-Instruct

LFM2.5

Liquid AI's 1.2B LFM2.5-Instruct with 262K downloads on HuggingFace. Extended 128K context for edge reasoning and instruction following.

Score 494 sources
textreasoningopen-sourceopen-weightself-hosted
Context
131,072
Input
Not applicable
Output
Not applicable
Coverage
Full profile

Liquid AI

LFM2.5-1.2B-Thinking

LFM2.5

Liquid AI's 1.2B LFM2.5-Thinking with chain-of-thought reasoning for edge devices. 30K downloads on HuggingFace.

Score 494 sources
textreasoningopen-sourceopen-weightself-hosted
Context
131,072
Input
Not applicable
Output
Not applicable
Coverage
Full profile