Use LiquidAI: LFM2.5-2.6B for free via OpenRouter
Liquid AI · 65,536 token context on this channel · Reasoning, Tool calling. This guide documents a specific API channel and has not run an inference request.
Check the free conditions first
Five steps to a first response
- 1
Create or sign in to your free account
Open OpenRouter and choose the free account. The published Free plan has API access and no payment method.
View Free plan ↗ - 2
Create an API key
Open the API Keys page while signed in and create a key. Copy it into a local environment variable; this site never asks for it.
Open API Keys ↗ - 3
Use this exact free variant
Copy the Base URL and Model ID below. Keep the :free suffix; the plain model ID can have different pricing.
- 4
Send the sample request from your terminal
Set OPENROUTER_API_KEY locally, then run the cURL example. It asks for a short text reply and does not enable paid fallback.
- 5
If the free limit is reached
Check your account limit or wait for the quota to reset. Recheck the official model page before trying another model; do not silently switch to a paid ID.
Check current limits ↗
Copy these exact public values
https://openrouter.ai/api/v1https://openrouter.ai/api/v1/chat/completionsliquid/lfm-2.5-2.6b:freeOpenAI Chat Completions · Bearer $OPENROUTER_API_KEYNone documented for this text requestMinimal cURL request
Run this on your own computer. Set OPENROUTER_API_KEY in your local shell first. The page never collects or sends your key. The sample requests at most 32 output tokens and contains no paid fallback.
# Set OPENROUTER_API_KEY in your own terminal first. Do not paste it into this site.
curl 'https://openrouter.ai/api/v1/chat/completions' \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H 'Content-Type: application/json' \
-d '{"model":"liquid/lfm-2.5-2.6b:free","messages":[{"role":"user","content":"Reply with one short greeting."}],"max_tokens":32}'A 401/403 can indicate a key or account issue; 404 can indicate a changed model ID; 429 usually points to a limit. Check OpenRouter's own error details before changing plans.
Where these details came from
Documentation checked 2026-09-29. Inference request: not run. Prices, quotas and access can change.
- OpenRouter model API ↗exact ID, zero prompt/completion price, modalities, context
- OpenRouter model page ↗current model and endpoint details
- OpenRouter Free plan ↗API access, no payment method, shared daily limit
- freeLLM model page ↗secondary model and phone-requirement cross-check
- freeLLM provider page ↗phone requirement is a third-party claim, not account-tested