AI Model Leaderboard 🚀

An interactive, developer-focused dashboard for benchmarking top free and commercial AI coding models. Designed for software engineers, security researchers, and developers to compare LLMs across performance scores, safety/refusal strictness, execution latency, and specialized developer workloads.

---

✨ Features

• Free vs. Commercial Tier Split: Instantly switch between free/open-weights models (GLM 5.3, DeepSeek V3/R1, Qwen 2.5 Coder) and paid API models (ChatGPT Pro/Plus, Claude 3.5 Sonnet, Gemini 2.0 Pro).
• Refusal Strictness Scale (0–100%): Clear safety index measuring guardrail strictness:
  • 10%–30% (Permissive): Ideal for penetration testing, shell scripting, and low-level dev ops.
  • 70%–100% (Strict): Enterprise guardrails with aggressive safety filters.
• Developer Workload Filters: View metrics specialized for:
  • ⚡ Cheat Coding & Speed-Run (Inline auto-complete & high tokens/sec)
  • 🛡️ Cybersecurity & Exploits (Vulnerability analysis & security tooling)
  • 🌐 Web APPs & Full-Stack (Multi-file web applications & backend code)
  • 🎨 UI & Design Systems (Tailwind CSS, React component visual fidelity)
  • 🏗️ Complex Architecture & Agents (Repository-level refactoring & long context)
• Interactive Metrics Chart: Built with Chart.js, supporting horizontal/vertical graph orientation and focal point filtering (Performance, Refusal Rate, Speed).
• Search & Multi-Metric Sorting: Real-time filtering by model name, provider, or technical features, with sorting by Performance, Speed, Strictness, or Context Window.
• Deep-Dive Modals: Click any model card to inspect technical specifications, pricing structures, context window limits, and sample benchmark prompts.

---

🛠️ Built With

• HTML5 & Vanilla JavaScript: Zero build steps or heavy framework dependencies.
• Tailwind CSS (CDN): Modern, enterprise dark-mode interface styled with slate/zinc tokens.
• Chart.js (CDN): Smooth animated responsive bar charts.
• Lucide Icons (CDN): Clean, lightweight icon set.

---

🚀 Quick Start / Deployment

Since the application is self-contained in a single file (index.html), deployment takes under 1 minute:

Option 1: GitHub Pages

1. Push index.html to your repository.
2. Go to Settings > Pages.
3. Select the main branch as the source and click Save.

Option 2: Local Execution

1. Clone the repository:

   git clone [https://github.com/failure10/ai-model-leaderboard.git](https://github.com/failure10/ai-model-leaderboard.git)

Open index.html in any web browser.

Made with ❤️ by failure10