> ## Documentation Index
> Fetch the complete documentation index at: https://docs.liquid.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Text Models

> Liquid's LFM text models range from 350M to 8B parameters, delivering ultra-low-latency generation while matching the performance of much larger models. They come in both dense and MoE variants to deploy flexibly across different devices.

<div className="use-cases" style={{marginTop: '-0.5rem'}}>
  <CardGroup cols={2}>
    <Card title="Chat & Rewriting" icon="font">
      Conversational interactions, text transformations, and summarization.
    </Card>

    <Card title="Tool Calling" icon="wrench">
      Function invocation, agent workflows, and API orchestration.
    </Card>

    <Card title="Structured Output" icon="brackets-curly">
      JSON generation, form filling, and data extraction from text.
    </Card>

    <Card title="Classification" icon="tags">
      Intent detection, routing, content labeling, and triage.
    </Card>
  </CardGroup>
</div>

<h2>LFM2.5 Models   <Badge shape="rounded" icon="circle-check" color="green" size="lg">Latest release</Badge></h2>

<Info>LFM2.5 builds on the LFM2 architecture with extended pre-training and reinforcement learning for improved chat, instruction-following, and tool-calling performance.</Info>

<CardGroup cols={2}>
  <Card title="LFM2.5-1.2B-Instruct" href="/lfm/models/lfm25-1.2b-instruct">
    1.2B · <Badge shape="pill" color="green">Recommended</Badge>

    Instruction-tuned for chat. Best for most use cases.
  </Card>

  <Card title="LFM2.5-1.2B-Thinking" href="/lfm/models/lfm25-1.2b-thinking">
    1.2B · <Badge shape="pill" color="purple">Reasoning</Badge>

    Optimized for math and logical problem-solving.
  </Card>

  <Card title="LFM2.5-1.2B-Base" href="/lfm/models/lfm25-1.2b-base">
    1.2B · <Badge shape="pill" color="orange">Pre-trained</Badge>

    Base model for finetuning or custom checkpoints.
  </Card>

  <Card title="LFM2.5-1.2B-JP" href="/lfm/models/lfm25-1.2b-jp">
    1.2B · <Badge shape="pill" color="red">Japanese</Badge>

    Fine-tuned model for high-quality Japanese text generation.
  </Card>

  <Card title="LFM2.5-350M" href="/lfm/models/lfm25-350m">
    350M · <Badge shape="pill" color="yellow">Fastest</Badge>

    Smallest LFM2.5 model for edge devices and low latency deployments.
  </Card>
</CardGroup>

## LFM2 Models

<CardGroup cols={3}>
  <Card title="LFM2-24B-A2B" href="/lfm/models/lfm2-24b-a2b">
    24B · 2B active · <Badge shape="pill" color="blue">MoE</Badge>

    Our largest model for laptops and single-GPU applications.
  </Card>

  <Card title="LFM2-8B-A1B" href="/lfm/models/lfm2-8b-a1b">
    8B · 1.5B active · <Badge shape="pill" color="blue">MoE</Badge>

    Mixture-of-experts model for on-device speed and quality.
  </Card>

  <Card title="LFM2-2.6B" href="/lfm/models/lfm2-2.6b">
    2.6B

    Highly capable model for deployment on most phones and laptops.
  </Card>

  <Card title="LFM2-2.6B-Exp" href="/lfm/models/lfm2-2.6b-exp">
    2.6B

    RL-only post-trained checkpoint for improved math and reasoning.
  </Card>

  <Card title="LFM2-1.2B" href="/lfm/models/lfm2-1.2b">
    1.2B · <Badge shape="pill" color="gray">Deprecated</Badge>

    Use the new LFM2.5-1.2B-Instruct checkpoint instead.
  </Card>

  <Card title="LFM2-700M" href="/lfm/models/lfm2-700m">
    700M

    Mid sized model for deploying on most devices.
  </Card>

  <Card title="LFM2-350M" href="/lfm/models/lfm2-350m">
    350M · <Badge shape="pill" color="yellow">Fastest</Badge>

    Our smallest model for edge devices and low latency deployments.
  </Card>
</CardGroup>

## Examples

Explore practical implementations using text models:

<CardGroup cols={2}>
  <Card title="Product Slogan Generator" href="/examples/android/slogan-generator">
    Generate creative marketing slogans on-device using LFM2-700M with single-turn generation and traditional Android Views.

    **Platform:** Android · **Uses:** LFM2-700M
  </Card>

  <Card title="Web Content Summarizer" href="/examples/android/web-content-summarizer">
    Share web pages from browsers to generate private, on-device summaries with web scraping and LFM2-700M processing.

    **Platform:** Android · **Uses:** LFM2-700M
  </Card>

  <Card title="Recipe Generator with Constrained Output" href="/examples/android/recipe-generator-constrained-output">
    Generate structured JSON recipes using LFM2-700M with constrained generation and the @Generatable annotation.

    **Platform:** Android · **Uses:** LFM2-700M
  </Card>

  <Card title="AI Agents with Koog Framework" href="/examples/android/leap-koog-agent">
    Build intelligent agents with tool invocation, MCP integration, and context management using LFM2-1.2B-Tool on Android.

    **Platform:** Android · **Uses:** LFM2-1.2B-Tool
  </Card>

  <Card title="English-Korean Translation CLI" href="/examples/laptop-examples/lfm2-english-to-korean">
    Bidirectional translation with automatic language detection. Fine-tuned LFM2-1.2B outperforms models 3x larger.

    **Platform:** Desktop · **Uses:** LFM2-1.2B (fine-tuned)
  </Card>
</CardGroup>

<Panel>
  * [Liquid Playground](https://playground.liquid.ai/chat?model=cmk0wefde000204jp2knb2qr8)
  * [HuggingFace Collections](https://huggingface.co/LiquidAI/collections)
  * [OpenRouter API](https://openrouter.ai/liquid)
  * [LEAP Model Library](https://leap.liquid.ai/models)
</Panel>