d1 Models by Liquid AI
Last refreshed 2026-10-29. Next refresh: weekly.
Details
Capabilities
About
d1 is Liquid AI's decision-model family: typed choice, yes/no (called "noul" in the API) or score answers with probabilities and zero output tokens. Hosted d1 (2026-09-22) is API-only on the Liquid Decisions API. Open-weight checkpoints d1-3B and d1-omni-600M landed on Hugging Face on 2026-10-07 under the LFM Open License v1.0. Not a chat LLM family.
Decision facts
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs classification, 128K context, and 1.9B parameters.
Use when the workload needs classification, 128K context, and 2.2B parameters.
Use when the workload needs structured outputs, multimodal inputs, and agent.
Use when the workload needs structured outputs, multimodal inputs, and agent.
Use when the workload needs 32K context, 3.1B parameters, and structured outputs.
Use when the workload needs 16K context, 587M parameters, and structured outputs.
Use when the workload needs structured outputs, multimodal inputs, and agent.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| d1-2B | Use when the workload needs classification, 128K context, and 1.9B parameters. | 2026-10 | classification128K context1.9B parameters | Current |
| d1-Omni-2B | Use when the workload needs classification, 128K context, and 2.2B parameters. | 2026-10 | classification128K context2.2B parameters | Current |
| d1.1 | Use when the workload needs structured outputs, multimodal inputs, and agent. | 2026-10 | structured outputsmultimodal inputsagent | Current |
| d1-ultra | Use when the workload needs structured outputs, multimodal inputs, and agent. | 2026-10 | structured outputsmultimodal inputsagent | Current |
| d1-3B | Use when the workload needs 32K context, 3.1B parameters, and structured outputs. | 2026-10 | 32K context3.1B parametersstructured outputs | Current |
| d1-omni-600M | Use when the workload needs 16K context, 587M parameters, and structured outputs. | 2026-10 | 16K context587M parametersstructured outputs | Current |
| d1 | Use when the workload needs structured outputs, multimodal inputs, and agent. | 2026-09 | structured outputsmultimodal inputsagent | Current |
Release Timeline
2 release groupsSpecifications(7 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Structured Outputs |
|---|---|---|---|---|---|---|
| d1-2B | 2026-10 | 128K | 1.92B | Yes | Yes | No |
| d1-Omni-2B | 2026-10 | 128K | 2.17B | Yes | Yes | No |
| d1.1 | 2026-10 | — | — | Yes | Yes | Yes |
| d1-ultra | 2026-10 | — | — | Yes | Yes | Yes |
| d1-3B | 2026-10 | 32K | 3.12B | Yes | Yes | Yes |
| d1-omni-600M | 2026-10 | 16K | 587M | Yes | Yes | Yes |
| d1 | 2026-09 | — | — | Yes | Yes | Yes |
Available From(3 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| d1 | Liquid AI | $0.02 | — | Serverless |
| d1 | Vercel AI Gateway | $0.02 | — | Serverless |
| d1 | OpenRouter | $0.02 | — | Serverless |
| d1.1 | Liquid AI | $0.02 | — | Serverless |
| d1.1 | OpenRouter | $0.02 | — | Serverless |
| d1-ultra | Liquid AI | $0.042 | — | Serverless |
| d1-ultra | Vercel AI Gateway | $0.042 | — | Serverless |
| d1-ultra | OpenRouter | $0.042 | — | Serverless |