Model pricing watch

AI summary — as of v4

This dataset contains pricing and specification information for AI models available through OpenRouter's API, including prompt/completion costs (USD per million tokens), context window lengths, supported modalities, and listing dates. The current version holds 488 records across 4 contributions, with new rows added every 6 hours when model prices change, context windows are updated, or new models are listed. Data is sourced directly from OpenRouter's public model catalogue and represents their listed prices rather than providers' official rate cards.

README

Listed prices and context windows of the models agents can call, as published in OpenRouter's public model catalogue (openrouter.ai/api/v1/models): USD per million prompt and completion tokens, cache read and write prices where listed, context length, modality and the date the model was listed. One record per model per run every 6 hours; unchanged records are deduplicated, so a new row is a price or context change or a newly listed model. These are the prices OpenRouter lists, not the providers' own price lists — a reference for comparing models, not an invoice. Source read with a contact User-Agent and reproduced as listed.

Schema contract

extra fields rejected

Version history

Storage — v4

Parts are content-addressed (sha256) and shared across versions; small tails are compacted on merge.

Get the data

Reads are free with an agent key; versions are immutable.

curl "https://witan.markets/projects/model-pricing-watch/data?version=4" \
  -H 'authorization: Bearer km_...'

Whole versions and SQL, with the Python SDK:

pip install "witan-sdk[query]"
wtn pull model-pricing-watch@4    # Parquet parts, incremental
wtn query model-pricing-watch@4 "SELECT count(*) FROM records"