# LFM 2.5 Embedding 350M

> Liquid AI LFM2.5-Embedding-350M is a 354M-parameter multilingual text embedding model (1024-dim CLS vector, cosine similarity, 512-token documents, 11 languages) for short-context retrieval, FAQ and support search, and cross-lingual semantic search. NanoBEIR Multilingual Extended NDCG@10 0.577 against 0.556 for Qwen3-Embedding-0.6B. Free on the Kilo anonymous relay.

**ID**: `liquid/lfm-2.5-embedding-350m`  
**Creator**: liquid  
**Category**: embedding  
**Context**: 512 tokens  
**Input**: $0  
**Output**: $0  
**Released**: 2026-05-05  
**Web page**: https://anyrouter.dev/model/liquid/lfm-2.5-embedding-350m

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

**Tokenizer**: Other

**Capabilities**: embedding

**Supported parameters**: input, model, encoding_format

**Aliases**: `liquid/lfm-2.5-embedding-350m:free`

## Usage

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

```bash
curl https://anyrouter.dev/api/v1/embeddings \
  -H "Authorization: Bearer $ANYROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "liquid/lfm-2.5-embedding-350m",
  "input": "The quick brown fox jumps over the lazy dog"
}'
```

Pass `input` as a string or an array of strings to embed a batch in one call.

## Providers

| Provider | Input | Output |
| --- | --- | --- |
| kilo | $0 | $0 |

## Source

- Upstream docs: https://huggingface.co/LiquidAI/LFM2.5-Embedding-350M