AI Model Parameters vs Training Dataset Size Analysis

Scatter plot showing the relationship between model parameter count and training dataset size

Both axes use logarithmic scale (log10) to handle the wide range of values

612
Total Models
15
Domains
10 - 2T
Parameter Range
4 - 30T
Dataset Size Range
Note: X-axis represents model parameters (log10 scale), Y-axis represents training dataset size in datapoints (log10 scale). Each point represents an AI model, colored by domain. Hover over points to see detailed information.

Scale Reference: X-axis: 1=10 params, 6=1M params, 9=1B params, 12=1T params | Y-axis: 3=1K datapoints, 6=1M datapoints, 9=1B datapoints, 12=1T datapoints