| name | free-ride |
| description | Free model fallback via OpenRouter. Use when: (1) primary API key is exhausted or rate-limited, (2) user wants to minimize costs, (3) background/low-priority tasks that don't need frontier models. Provides a ranked list of free models on OpenRouter sorted by capability. |
Free Ride — OpenRouter Free Model Fallback
Use OpenRouter's free model tier as a fallback when paid APIs are unavailable or for low-priority tasks.
Setup
- Get a free API key at https://openrouter.ai/keys
- Add to your environment (already configured in
~/.openclaw/.env):
OPENROUTER_API_KEY=sk-or-v1-xxxxx
Status: CONFIGURED
Free Models (ranked by capability, March 2026)
| Model | Context | Speed | Best For |
|---|
deepseek/deepseek-chat:free | 128k | Fast | General chat, Chinese |
google/gemma-3-27b-it:free | 96k | Fast | Reasoning, multilingual |
mistralai/mistral-small-3.1-24b-instruct:free | 96k | Fast | Code, instruction following |
qwen/qwen3-32b:free | 40k | Medium | Chinese, general |
meta-llama/llama-4-scout:free | 512k | Medium | Long context |
google/gemini-2.5-pro-exp-03-25:free | 1M | Slow | Complex reasoning |
OpenRouter Config for OpenClaw
Add as a provider in openclaw.json (manual edit recommended):
{
"models": {
"providers": {
"openrouter_free": {
"baseUrl": "https://openrouter.ai/api/v1",
"apiKey": "YOUR_OPENROUTER_KEY",
"api": "openai-completions",
"models": [
{
"id": "deepseek/deepseek-chat:free",
"name": "DeepSeek Chat (Free)",
"api": "openai-completions",
"reasoning": false,
"input": ["text"],
"cost": {"input": 0,
Usage Strategy
- Add
openrouter_free/deepseek-chat:free as a fallback in agent model config
- Use for: background cron tasks, routine scans, simple classification
- Don't use for: complex reasoning, code generation, production-critical tasks
- Rate limits apply per model — spread load across multiple free models
Warning
Do NOT let OpenClaw auto-install this skill's config changes. Edit openclaw.json manually or use a coding agent to avoid breaking your existing provider setup.