usage
Model usage
Fourteen completed UTC days are split into previous and current seven-day windows. Model values are observed floors from days in the daily Top 50, with Other retained in the denominator.
Methodology and data sources
How LastAI validates, normalizes, and presents the source data. Source information is disclosed without linking to an external service.
usage-filters
Context · tools · images
The same 14-day dataset is filtered for the 10K context bucket, requests containing tool calls, and image-input activity. Values are observed token floors, not request, tool-call, or image counts.
tasks
Task classification
Request and token shares come from a classified sample and are not interpreted as absolute traffic or total market demand.
apps
Apps · agents
Only apps opting into public attribution are included. Placements are joined by source App ID, and official URLs are never inferred.
benchmarks
Benchmarks
Third-party metrics retain their source meaning and are not blended into usage ranking or a LastAI composite score.
weekly
Weekly report
Monday's sync (last week, Monday to Sunday, and the week before) is kept as a weekly record that is never changed. Model and maker shares compare the two weeks in that record, app ranks compare with last week's record, and the result is published at /rankings/weekly/usage.
ATTRIBUTION
Original data source
Source: OpenRouter. LastAI is not the original provider and presents a separate analysis based on documented API responses at collection time.
Models APIopenrouter.ai
Daily rankings datasetopenrouter.ai
Task classifications datasetopenrouter.ai
App rankings datasetopenrouter.ai
Benchmarks datasetopenrouter.ai
Snapshot as of 2026-11-12T05:20:08.915Z · Synced 2026-11-12T05:20:14.971Z