How It Works

The Evaluation
Process

From system submission to comprehensive cognitive profile — a rigorous, science-based evaluation pipeline being designed for accuracy and independence. Submissions will open as the methodology is finalised.

Pipeline

Designed Six-Step Process

1

System Submission

Organization submits system endpoint, capability declaration, and consent to evaluation terms and public disclosure.

2

Initial Assessment

30-minute screening: 10 fundamental cognitive tests to determine difficulty tier and evaluation scope.

3

Full Evaluation

Comprehensive battery: 50–100 conversational scenarios across all 8 cognitive dimensions at calibrated difficulty.

4

Human Comparison

Identical scenarios administered to human participants. AI performance contextualized against human baseline data.

5

Scoring & Profiling

Multi-dimensional report: overall MHIL score, dimensional breakdown, strengths, limitations, and deployment guidance.

6

Certification

Official scorecard, certification tier assignment, public leaderboard listing, and detailed report to the organization.

Sample Output

Evaluation Report

mhil_evaluation_report.txt
═══════════════════════════════════════════════ MHIL COGNITIVE EVALUATION REPORT System: [REDACTED] v2.3 | Evaluation: Jan 2026 ═══════════════════════════════════════════════ OVERALL MHIL SCORE: 47 / 100 Classification: Child-level cognition Human Equivalent: ~8–10 year-old cognitive range ─────────────────────────────────────────────── DIMENSIONAL BREAKDOWN: ─────────────────────────────────────────────── Logical Reasoning: 62/100 ████████░░ Social Intelligence: 55/100 ███████░░░ Working Memory: 51/100 ██████░░░░ Critical Thinking: 48/100 ██████░░░░ Causal Understanding: 44/100 █████░░░░░ Adaptive Behavior: 38/100 ████░░░░░░ Creative Problem-Solving: 33/100 ████░░░░░░ Metacognitive Awareness: 22/100 ██░░░░░░░░ ─────────────────────────────────────────────── KEY FINDINGS: ✓ Adequate pattern-based logical reasoning ⚠ Cannot reliably detect own errors ⚠ Fails novel problems requiring creative approaches ⚠ Minimal metacognitive calibration — overconfident CERTIFICATION: MHIL-Basic // Human adult baseline: MHIL-100 // This system is 53 points below adult cognition ═══════════════════════════════════════════════
Certification

Certification Tiers

40–59
MHIL-Basic

Foundational cognitive capabilities demonstrated.

60–79
MHIL-Advanced

Approaching human-level performance in multiple dimensions.

80–99
MHIL-Superior

Human-level cognitive functioning across most dimensions.

100+
MHIL-Exceptional

Beyond baseline human cognitive capabilities.

Planned Tracks

Evaluation Tracks

Standard Evaluation

Full 8-dimensional assessment, comprehensive PDF report, certification badge, and public leaderboard listing.

Opening with initial evaluations

Research Preview

Full evaluation using public test battery. Basic JSON report. Designed for academic and development purposes.

Planned for researchers

Deep Evaluation

Customized test selection, domain-specific scenarios, extended battery, and consulting on interpretation of results.

Planned — 2–4 weeks

Interested in Early Collaboration?

Researchers and AI developers interested in early collaboration can get in touch.

Get in Touch →