AI Market Pulse
A single number that captures how much it costs to get one unit of quality from an AI API. We take each model's average token price and divide it by its benchmark score, then report the median across all scored models, scaled to a readable range.
Across 124 models with quality scores
Lower = Cheaper AI
When the index falls, you get more capability per dollar. Price cuts and new efficient models push it down.
How It's Calculated
For each model: average token price ÷ quality score. The TPI is the median of those ratios, scaled × 10 000.
Tiers Show Segments
Budget, Mid-Range, and Premium tiers track separate sub-indices so you can see where prices are moving in your price band.
Budget (<$1/M) · Mid-Range ($1–$10/M) · Premium (>$10/M)
Budget
70 models
Mid-Range
48 models
Premium
6 models