Last refreshed 2026-09-29. Next refresh: weekly.
Why use d1 on Liquid AI?
Liquid AI offers d1 with free input token pricing. Liquid AI provides its LFM models through the LEAP on-device inference SDK and LEAP Model Library.
Setup recipe
Docs fallbackUse the provider REST API or SDKCreate a provider API keymodel: d1:freed1:freeRequest example
d1:free.Gotchas
- Use provider model ID "d1:free", not the LLMReference slug "d1".
Pricing
| Type | Price (per 1M) |
|---|---|
| Input tokens | Free |
| Output tokens | Free |
Capabilities
About d1
d1 (API id d1:free) is Liquid AI's first decision model: structured Noul/Choice/Score decisions with calibrated probabilities and zero generated output tokens via POST https://api.liquid.ai/decisions/v1/systemone (TypeSafe SDK–compatible). First-party docs (decision-models) + GET /decisions/v1/models catalog: release_date 2026-09-22; complete-library matrix Family=d1, distribution API-only (HF/GGUF/MLX/ONNX unavailable), not trainable. Tip @liquidai 2026-09-29: multilingual evals, prompt-injection robustness, longer inputs, OpenRouter soon. No first-party context length or $/1M rate card published this pass — pmid d1:free implies free tier ($0/$0 on liquid join). Proprietary hosted weights. Peer decision model under its own family d1 (researcher liquid-ai); do not attach under system-one, jev, lfm-2, or lfm-2.5.