d1-Omni-2B

Use it to classify or extract information from text and media, locate objects in images, estimate quantities, or make decisions based on audio such as routing calls.

Availability

Checkpoint available on Hugging Face; API and playground links provided

Pricing model

Not established

Deployment

Runs in Transformers

Privacy

Not established

Licence

LFM Open License v1.0

• The model returns typed answers with probabilities in a single forward pass and generates no tokens. ^[1]
• Inputs can include text or JSON, pictures, video frames, audio clips, and written or spoken questions. ^[1]

  <td style="border: 0; padding: 0.55em 0.75em;">A state (text or JSON), pictures, a video, audio clips, and any number of named questions, written or spoken</td>

• It can extract text, numbers, and dates with spans, locate objects in pictures, and estimate quantities with an interval. ^[1]

  - **Beyond classification:** it also extracts text, numbers, and dates with their span, locates objects in pictures, and estimates quantities with an interval.

• The model checkpoint is available as LiquidAI/d1-Omni-2B on Hugging Face, with Transformers inference instructions. ^[1]
• The model runs in Transformers with its own model code. ^[1]

  d1-Omni-2B runs in [Transformers](https://docs.liquid.ai/deployment/gpu-inference/transformers) with its own model code.

• The model is designed for decisions based on audio, including call and voice-message routing and sound events. ^[1]

  d1-Omni-2B is designed for decisions that depend on what is heard, such as call and voice-message routing and sound events, and it handles the same text and picture decisions as d1-2B.

• New models bring live speech translation and voice features Daily brief 2026-10-28