# LFM3 1.7B

> LFM3 1.7B is Liquid AI's 1.73B-parameter text model for local agents, tool calling and structured outputs. It has a 256K-token context window and reasons before it answers. It is open weight and small enough to run on a phone or laptop.

**ID**: `liquid/lfm-3-1.7b`  
**Creator**: liquid  
**Category**: text  
**Context**: 262K tokens  
**Pricing**: $0.08 / $0.20 per 1M · billed by your provider (BYOK)  
**Released**: 2026-10-27  
**Links**: [Weights on Hugging Face](https://huggingface.co/LiquidAI/LFM3-1.7B) · [Announcement](https://www.liquid.ai/blog/lfm3) · [AnyRouter release](https://anyrouter.dev/blog/releases/2026-11-15-lfm-3-1-7b)  
**Web page**: https://anyrouter.dev/model/liquid/lfm-3-1.7b

**Input modalities**: text  
**Output modalities**: text  

**Tokenizer**: Other

**Capabilities**: chat, streaming, reasoning, function-calling

**Aliases**: `liquid/lfm3-1.7b`

## Usage

**Endpoint**: `POST https://anyrouter.dev/api/v1/chat/completions`

```bash
curl https://anyrouter.dev/api/v1/chat/completions \
  -H "Authorization: Bearer $ANYROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "liquid/lfm-3-1.7b",
  "messages": [
    {
      "role": "user",
      "content": "Say hi in 3 words."
    }
  ]
}'
```

## Providers

| Provider | Input | Output |
| --- | --- | --- |
| openrouter-byok | $0.08/M (list) | $0.20/M (list) |
| nousresearch-byok | $0.08/M (list) | $0.20/M (list) |
| infron-byok | $0.08/M (list) | $0.20/M (list) |
| nano-gpt-byok | $0.08/M (list) | $0.20/M (list) |

BYOK routes are billed by your own provider key — AnyRouter charges $0. Rates marked (list) are the provider's published list price.

## Model card

- Total parameters: 1.73B [2,1]
- Native context: 262144 tokens [2,1]
- License: LFM Open License v1.0, open weights (https://huggingface.co/LiquidAI/LFM3-1.7B) [2]

### Strengths [1,2]

- Outperforms Liquid's previous LFM2.5-2.6B while using 35% fewer parameters.
- Fits in about 1 GB of memory at 4-bit precision, so it runs on phones and laptops.
- Post-trained for agentic harnesses, tool calling and structured outputs, with a 256K-token context window.

### Known limits [2]

- Not recommended for agentic coding or complex multi-step reasoning because of its small size.
- Use retrieval or search for knowledge-heavy tasks.

### Use cases [1,2]

- Local agents and tool calling on edge devices.
- Structured output extraction.
- A small base for single-GPU fine-tuning on a specialised task.

### Links

- announcement: https://www.liquid.ai/blog/lfm3
- model card: https://huggingface.co/LiquidAI/LFM3-1.7B
- docs: https://docs.liquid.ai/lfm/getting-started/welcome
- Previous version: https://anyrouter.dev/model/liquid/lfm-2.5-2.6b

### Sources

1. [Introducing LFM3: Every Modality for Specialized Intelligence](https://www.liquid.ai/blog/lfm3) (accessed 2026-11-15)
2. [LiquidAI/LFM3-1.7B model card](https://huggingface.co/LiquidAI/LFM3-1.7B) (accessed 2026-11-15)

## Source

- Upstream docs: https://www.liquid.ai/blog/lfm3