AI model selection & cost estimator for Zo Computer
Choose the best AI model based on task, budget, speed, and context size. Get instant recommendations with detailed comparisons and cost projections.
- 🎯 Smart Recommendations — Scored model matching based on your criteria
- 💰 Token Cost Estimator — Per-request cost with 1K/10K/100K projections
- 📊 Sortable Comparison Table — All 8 models side-by-side with expandable details
- 🌓 Dark/Light Mode — Toggle with smooth transitions
- 📥 Export to Markdown — Download a full report
- ⚡ Quick Presets — Chat turn, Code gen, RAG query, Doc summary, Full codebase
| Model | Provider | Input/1M | Output/1M | Context |
|---|---|---|---|---|
| GPT-5.4 mini | OpenAI | $0.75 | $4.50 | 400K |
| GPT-5.4 | OpenAI | $2.50 | $10.00 | 272K |
| GPT-5.3 Codex | OpenAI | $1.75 | $14.00 | 400K |
| Opus 4.6 | Anthropic | $5.00 | $25.00 | 1M |
| Sonnet 4.5 | Anthropic | $3.00 | $15.00 | 200K |
| Gemini 3 Pro | $2.00 | $12.00 | 200K | |
| Kimi K2.5 | Moonshot AI | $0.60 | $3.00 | 262K |
| MiniMax 2.7 | MiniMax | $0.30 | $1.20 | 1M |
bun run Skills/model-advisor/scripts/install.ts# Clone into your Skills directory
cd Skills
git clone https://github.com/KaiyzerBX50/model-advisor.git
# Deploy to your zo.space
bun run model-advisor/scripts/install.ts--path <path> Route path (default: /model-advisor)
--private Make the page private (owner-only)
--help Show help
"Install the model-advisor skill to my space"
- Runtime: Zo Space (React + Tailwind CSS 4)
- Dependencies: None (uses pre-installed zo.space packages)
- Icons: lucide-react
- State: React hooks (useState, useMemo, useCallback)
- Theming: Context-based dark/light mode
model-advisor/
├── SKILL.md # Zo skill manifest
├── README.md # This file
├── LICENSE # MIT License
├── assets/
│ └── model-advisor.tsx # Full page source code
└── scripts/
└── install.ts # One-click installer
Edit assets/model-advisor.tsx to:
- Add/remove models in the
MODELSarray - Adjust scoring weights in
scoreModel() - Modify task categories in
TASKS - Update pricing data
Then re-run the install script to redeploy.
- MythologIQ — Major refactor: decomposed monolith into 11 modules, fixed 6 data errors, added benchmark-based scoring, +4 new models
- DaGAWD NYC — Original creator and project lead
MIT — use it however you want.
Built with Zo Computer ⢕⣿⣏⣿⢳⡕⣆⢺⣋⢟