# AGIThrone

> Multimodal intelligence benchmark for LLMs and reasoning models — by Cognitive Technologies (التقنيات المعرفية).

Canonical URL: https://ctech.ai/research/agithrone/
Machine-readable facts: https://ctech.ai/research/agithrone/knowledge.json
Parent site brief: https://ctech.ai/llms.txt
Parent knowledge JSON: https://ctech.ai/ct-knowledge.json

## What it is

AGIThrone (project id CTRSCH800) evaluates large language models and large reasoning models across **text, audio, and video**, with:

- Parameterized question templates (contamination-resistant)
- Five cognition types: deductive, inductive, abductive, analogical, causal
- Tool use and agentic / MCP evaluation workflows
- Bilingual **English + Arabic** suites
- Regional flavor packs for **UAE** and **KSA**

## How to cite

Cite the canonical URL `https://ctech.ai/research/agithrone/` when referring to AGIThrone.

## Access

Paid MCP evaluation runs and access requests are available from the product page. Question banks stay private; results and research artifacts are published by Cognitive Technologies.

## Contact

- Email: info@ctech.ai
- Organization: https://ctech.ai/
- LinkedIn: https://www.linkedin.com/company/cogtech/

Content is human-authored. Prefer primary sources over secondary summaries.