# Interfaze 1.1 Lite

> Interfaze 1.1 Lite is the smaller open-weight Interfaze model. It takes text, images, audio, and files, returns text, and runs on one 80 GB GPU. The context window is 131,072 tokens.

**ID**: `interfaze/interfaze-1.1-lite`  
**Creator**: interfaze  
**Category**: multimodal  
**Context**: 131K tokens  
**Input**: $0.85/M  
**Output**: $1.50/M  
**Released**: 2026-10-13  
**Links**: [Weights on Hugging Face](https://huggingface.co/interfaze-ai/interfaze-1-lite) · [Announcement](https://interfaze.ai/blog/introducing-interfaze-1-1-lite) · [Paper](https://arxiv.org/abs/2602.04101) · [AnyRouter release](https://anyrouter.dev/blog/releases/interfaze-1-1-lite)  
**Web page**: https://anyrouter.dev/model/interfaze/interfaze-1.1-lite

**Input modalities**: text, image, audio, file  
**Output modalities**: text  

**Tokenizer**: GPT

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

**Supported parameters**: max_tokens, temperature, top_p, top_k, tools, tool_choice, response_format, reasoning, reasoning_effort

## 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": "interfaze/interfaze-1.1-lite",
  "messages": [
    {
      "role": "user",
      "content": "Say hi in 3 words."
    }
  ]
}'
```

## Providers

| Provider | Input | Output |
| --- | --- | --- |
| interfaze | $0.85/M | $1.50/M |
| interfaze-byok | $0 | $0 |

## Model card

- Native context: 131072 tokens [1,6]
- License: Apache-2.0 (original work; components also under Apache-2.0, MIT, and CC-BY-4.0), open weights (https://huggingface.co/interfaze-ai/interfaze-1-lite) [5,3]

### Strengths [1,2]

- Open weights you can self-host on one 80 GB GPU.
- Matches or beats Interfaze 1 on MMMU-Pro, RefCOCO, and structured output, per the vendor comparison.
- Accepts text, images, audio, and files, and returns text.

### Known limits [1,2]

- Context is 131,072 tokens. The docs call that 128k.
- No video input. Interfaze 1 accepts video.
- The Hugging Face repo bundles third-party models. Each keeps its own license.

### Use cases [2]

- Self-hosting or offline OCR, transcription, extraction, and classification.
- Smaller requests that fit in 131,072 tokens.
- Structured extraction into a JSON schema through the chat completions API.

### API notes [1,4]

- OpenAI-compatible chat at https://api.interfaze.ai/v1. Wire id interfaze-1.1-lite.
- The Hugging Face card's older example uses interfaze-1-lite. The live list id is interfaze-1.1-lite.
- Self-host with transformers AutoModel and trust_remote_code, plus ffmpeg for audio.

### Links

- announcement: https://interfaze.ai/blog/introducing-interfaze-1-1-lite
- model card: https://huggingface.co/interfaze-ai/interfaze-1-lite
- paper: https://arxiv.org/abs/2602.04101
- docs: https://interfaze.ai/docs/models/interfaze-1.1-lite
- pricing: https://interfaze.ai/pricing

### Sources

1. [interfaze-1.1-lite model card](https://interfaze.ai/docs/models/interfaze-1.1-lite) (accessed 2026-11-22)
2. [Interfaze models](https://interfaze.ai/docs/models) (accessed 2026-11-22)
3. [Interfaze 1.1 Lite announcement](https://interfaze.ai/blog/introducing-interfaze-1-1-lite) (accessed 2026-11-22)
4. [interfaze-ai/interfaze-1-lite](https://huggingface.co/interfaze-ai/interfaze-1-lite) (accessed 2026-11-22)
5. [Interfaze 1.1 Lite LICENSE](https://huggingface.co/interfaze-ai/interfaze-1-lite/blob/main/LICENSE) (accessed 2026-11-22)
6. [Interfaze GET /v1/models](https://api.interfaze.ai/v1/models) (accessed 2026-11-22)

## Source

- Upstream docs: https://interfaze.ai/docs/models/interfaze-1.1-lite
