LFM2.5 350M

Name: LFM2.5 350M
Author: Liquid AI

by Liquid AILFM2.5

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released November 2025128K context≈ 96,000 words

A remarkably compact model that punches above its weight class for its size. LiquidAI's LFM2.5 350M uses a hybrid liquid foundation architecture rather than a standard transformer, giving it efficient handling of long contexts despite having only 350 million parameters. It trades raw reasoning depth for speed and a surprisingly large 128k context window at minimal compute cost.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Creative Writing

Moderate

Long Context

LFM2.5 350M

by Liquid AILFM2.5

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released November 2025128K context≈ 96,000 words

A remarkably compact model that punches above its weight class for its size. LiquidAI's LFM2.5 350M uses a hybrid liquid foundation architecture rather than a standard transformer, giving it efficient handling of long contexts despite having only 350 million parameters. It trades raw reasoning depth for speed and a surprisingly large 128k context window at minimal compute cost.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Creative Writing

Moderate

Long Context

Glossary

ArchitectureThe underlying structural design of a neural network that defines how data flows through layers and components.Compact ModelA smaller language model designed to use fewer computational resources while still performing useful tasks.Context WindowThe maximum number of tokens a model can process in a single conversation or prompt.Liquid Foundation ArchitectureA neural network design that differs from standard transformers by using continuous, adaptive mechanisms instead of fixed attention patterns to process information.ParametersThe learned numerical values in a model — more parameters generally means more capacity but higher compute cost.ReasoningThe model's ability to work through multi-step logical problems and provide justified answers rather than just pattern-matching.Reasoning DepthA model's ability to perform complex multi-step logical thinking and problem-solving; typically increases with model size.TransformerThe dominant neural network architecture for language models, using self-attention to process sequences.

Capabilities

Capabilities

Use Case Fit

Glossary