AI search intelligence layer
Artificial Search
Find the best AI model, agent, MCP server and skill pack for your task.
Recommended stack preview
Prototype keyword matching from demo data
Task
Fix bugs
Model
Claude Sonnet / GPT-5.5
Agent
Codex or Claude Code
Skill Pack
Fullstack Debugging Pack
Prototype leaders
Demo leaders across models, agents, MCP and movement signals. Not real-time yet.
Demo Intelligence Index Over Time
Mock weekly trend, normalized to a 0-100 demo index.
Top AI Models in Prototype Scoring
Ranked by the demo scoring formula from the task file.
| Rank | Model | Provider | Intelligence | Coding | Agent Power | Speed | Cost Efficiency | Context | Overall Score |
|---|---|---|---|---|---|---|---|---|---|
| #1 | GPT-5.5 | OpenAI | 96 | 95 | 94 | 78 | 72 | 128K tokens | 89.9 |
| #2 | Claude Opus / Sonnet | Anthropic | 95 | 93 | 91 | 82 | 76 | 200K tokens | 89.6 |
| #3 | DeepSeek | DeepSeek | 88 | 90 | 82 | 84 | 93 | 128K tokens | 87.7 |
| #4 | Gemini | 91 | 86 | 84 | 88 | 80 | 1M tokens | 86.4 | |
| #5 | Qwen | Alibaba | 86 | 88 | 80 | 86 | 89 | 128K tokens | 85.7 |
| #6 | Kimi | Moonshot AI | 87 | 84 | 81 | 83 | 86 | 1M tokens | 84.5 |
| #7 | Mistral | Mistral AI | 84 | 82 | 78 | 89 | 88 | 128K tokens | 83.4 |
| #8 | Llama | Meta / local hosts | 80 | 79 | 72 | 74 | 95 | 32K tokens | 79.8 |
Not just a leaderboard
Artificial Search compares models, AI agents, MCP servers, skill packs and concrete workflows. A model can be strong in reasoning but weak as an agent host; an agent can use tools well but need a safer MCP permission profile; a workflow can need a focused pack rather than a generic prompt.
Models
Demo scoring, filters and practical fit signals.
Agents
Demo scoring, filters and practical fit signals.
MCP risk
Demo scoring, filters and practical fit signals.
Skill packs
Demo scoring, filters and practical fit signals.