model-intel

What This Skill Does

Model Intel is an essential command-line utility for OpenClaw agents, providing real-time intelligence on LLM availability, pricing, and performance metrics via the OpenRouter API. By integrating this skill, your agent can instantly query live data to make informed decisions regarding model selection. Whether you are optimizing for operational costs, latency, or high-level reasoning capabilities, Model Intel acts as a source of truth for the rapidly evolving landscape of Large Language Models. It eliminates the guesswork associated with model selection by providing side-by-side comparison tools and categorization for specific use cases like code generation, vision, and long-context analysis.

Installation

To integrate this skill into your OpenClaw environment, ensure you have the OpenClaw CLI properly configured, then run the following command in your terminal:

clawhub install openclaw/skills/skills/aiwithabidi/model-intel-pro

Once the installation completes, the agent will have immediate access to the model_intel.py script. Please verify that your system environment has the necessary network permissions to reach the OpenRouter API endpoint to ensure the data remains live and up-to-date.

Use Cases

The Model Intel skill is specifically designed for developers and power users who build production-grade AI workflows.

• Budget Optimization: Use the best cheap command to find the most cost-effective models without compromising necessary performance for high-volume tasks.
• Performance Tuning: Use best fast to identify models with low time-to-first-token for real-time chat interfaces.
• Complex Problem Solving: Use best reasoning to select models capable of solving complex math, logic, or architectural tasks.
• Resource Management: Quickly lookup technical specifications like context windows and per-million-token pricing for any supported model.

Example Prompts

1. "Check the current pricing for Claude-3.5-Sonnet and compare it to GPT-4o-mini to see which is cheaper for our batch processing task."
2. "I need to analyze a 50,000-word document. What is the best model available right now that supports a large context window?"
3. "List the top-rated models specifically for Python coding tasks and show me the price difference between them."

Tips & Limitations

• Live Data: Always use the command-line utility to pull the latest stats; avoid caching results for more than 24 hours as model pricing and availability fluctuate daily.
• Scope: This skill is focused specifically on OpenRouter-supported models; it does not cover private self-hosted instances.
• Context: When using the best command, provide as much detail as possible about your hardware or latency requirements to get the most accurate recommendations.