OG Stress Tester — OpenGradient Model Quality Analyser

OpenGradient Model Stress Tester

Paste any model CID from the OpenGradient Hub. The tester generates 100 input vectors — 96 random samples across the full input range plus 4 designed edge cases — runs them all, and returns a full statistical analysis of your model's output distribution with a TEE-verified AI verdict.

Model CID (from Hub)

Input Profile

Quick load:

ETH Volatility

DeFi Liquidation

Credit Score

Wallet Fraud

Token Direction

// Custom Input Ranges

Running stress test 0 / 100

Generating 100 input vectors

Running batch inference on OpenGradient

Computing output statistics

Getting TEE-verified AI verdict

Settling proof on OpenGradient Network

// TEE-Verified AI Verdict · OpenGradient Network

Verdict loading...

ON-CHAIN PROOF

—

// Samples Run

—

+ edge cases

// Failure Rate

—

inference errors

// Output Range

—

min → max

// Consistency

—

/ 100

// Output Distribution Histogram —

— — —

// Percentile Distribution

Percentile	Value	Interpretation
P5	—	Extreme low outputs
P25	—	Lower quartile
P50 (Median)	—	Middle value
P75	—	Upper quartile
P95	—	Extreme high outputs
Mean	—	Average output
Std Dev	—	Output spread

// Model Quality Assessment

Metric	Value
Edge Case Sensitivity	—
Edge Case Mean Output	—
Random Sample Mean	—
Distribution Shape	—
Output Stability	—

// Sample Inputs & Outputs 🟡 = edge case

#	Type	First 3 Inputs	Output