The Falling Cost of Intelligence
X = release date. Y = effective cost per 1M I/O Tokens (log). Each step line tracks the cheapest model to date at or above an IQ level; large dots mark the models that set each new price floor, faint dots show every other qualifying model, colored by provider.
The Falling Cost of Intelligence
CoreIQ Over TimeEach model's estimated IQ by release date. Step-lines connect each provider's flagship models over time.Data: ARC Prize, Epoch AI, Epoch AI +26 moreOpen chartCostIQ vs CostEach model's estimated IQ against its effective cost per 1M I/O Tokens (sticker price × measured or imputed usage multiplier).Data: AI IQ methodologyOpen chartCostAI Models by CostPublished price for 1M I/O Tokens compared with effective cost after applying the measured or imputed task-usage multiplier.Data: Artificial Analysis, ARC Prize, Vals Index v2 Cost per Test +1 moreOpen chartCostTask EfficiencyEach dot shows the inverse of the effective-cost usage multiplier. Higher means less price-adjusted task work: 2× is about half the median task effort. Source-backed multipliers are preferred; lineage, peer, and 1× fallbacks are labeled in tooltips.Data: Artificial Analysis, ARC Prize, Vals Index v2 Cost per Test +1 moreOpen chartCostInput Price vs Output PriceEach model positioned by its published token prices — input price (Y) against output price (X), both per 1M tokens on a log scale. The dashed line is the best fit across models, showing the typical output-to-input price relationship.Data: AI IQ datasetOpen chartCostIQ vs Input/Output Token CostEach model's estimated IQ plotted against its published token price. Toggle between input and output price per 1M tokens.Data: AI IQ methodologyOpen chart