Philanthropiaic Efficiency Index

The standard for AI performance auditing.

We focus on practical, measurable performance. We audit industry models based on real energy consumption, accuracy, and operational speed.

Live Performance Data Active Node
Model FEGE Score Accuracy Speed (TPS) Energy Cost / Run
Retrieving latest benchmark data...

Phase 1 Telemetry Calibration: Hardware energy (Joules) and dynamic API cost pipelines are undergoing rigorous engineering calibration. Uncalibrated mathematical variables safely default to 0 to prevent data hallucination.

Strict 45s Timeout Cutoff: Timeouts exceeding 45 seconds are classified as operational failures. These models are immediately penalized with a punitive score of 0 for both Accuracy and FEGE Index.

Pending Models in Queue

Metrics Glossary

FEGE Efficiency Score

Our proprietary metric that balances accuracy, speed, and energy consumption. A higher score represents better overall optimization.

Accuracy

Evaluated via deterministic string distance logic against strict expected outputs. We do not use other AI models to judge results.

Speed (TPS)

Tokens Per Second generated dynamically. Measures the actual processing speed of the underlying hardware.

Energy Consumption

Estimated physical energy in Joules required to process the test, calculated using the Thermal Design Power (TDP) of standard hardware.

The Unified Formula

Our framework evaluates the balance between cognitive output and hardware strain:

FEGE = (Accuracy * (TPS / 10)) / ((Config + 1) * (1 + Variance) * (Joules / 1000) * (Cost_USD * 100) + 1)