Artificial Analysis Skill

artificial-analysis-skill

An agent skill that picks the right AI model for any task using live benchmark data from [artificialanalysis.ai](https://artificialanalysis.ai) — intelligence index, coding, math, output speed, time-to-first-token, pricing, and context window for 400+ LLMs, plus ELO leaderboards for text-to-image, image-editing, text-to-speech, text-to-video, and image-to-video models.

Works with any agent that supports the [skills.sh](https://skills.sh) format — Claude Code, Cursor, Codex, Cline, Gemini CLI, Copilot, and more.

Why

Model rankings change weekly. Your coding agent's training data is stale. This skill teaches it to **always check the live leaderboard** before answering "which model should I use?" and to pick with a weighted-scoring rubric instead of vibes.

Install

npx skills add mrloldev/artificial-analysis-skill

That's it. The skill is auto-discovered by your agent.

Manual install

Clone into your agent's skills directory:

git clone https://github.com/mrloldev/artificial-analysis-skill \
  ~/.claude/skills/artificial-analysis

Works equivalently under `.cursor/skills/`, `.agents/skills/`, `.codex/skills/`, etc.

Setup

1. Get a free API key (1,000 requests/day) at [artificialanalysis.ai/documentation](https://artificialanalysis.ai/documentation).

2. Export it:

   export AA_API_KEY=your_key_here

3. Install dependencies: `curl` and `jq`.

   brew install curl jq          # macOS
   sudo apt install curl jq      # Debian/Ubuntu

Responses are cached at `~/.cache/artificial-analysis/` for 24h, so repeated queries cost zero rate-limit.

Usage

Once installed, just ask your agent in plain English:

"What's the best LLM right now for agentic coding under $10/M tokens?"

"Compare claude-4-5-sonnet with gpt-5-xhigh and gemini-3-1-pro."

"Cheapest model that scores above 60 on intelligence index?"

"Which text-to-video model is ranked highest?"

The skil