How we source model data

Prices, context windows and lifecycle dates in our AI tools come from each vendor's official pricing, model and deprecation pages. Every catalog row records the page it was read from and the date it was checked.

Current models: 39; from 9 providers
Prices verified: October 3, 2026; against vendor pricing pages
Sources scanned: October 3, 2026; 95 official sources

How the catalog is maintained

Primary sources only. Rates are read from Anthropic, DeepSeek, Google, Groq, Meta, Mistral, OpenAI, Qwen, xAI directly, not from aggregators or press coverage.

Weekly checks. Vendor pages are scanned every week. A changed price or a new model is extracted, cross-checked independently, and only then published.

Estimates and editorial judgments

Scores and recommendations in the AI Model Picker are our editorial assessment. Embedding limits, hardware sizing and latency figures in the planning calculators are estimates for capacity planning, not benchmarks.

For contracts, compliance reviews or production budgets, confirm the figures with the vendor.

Browse and integrate

/models/

Search models by author, host and modality

/llm-pricing-api/

Model prices and limits as JSON and CSV

/api/model-candidates.json

Paginated model directory API

/api/data-quality.json

Source coverage and check dates

Spotted an error?

Email hello@aidevhub.io with the model name and the vendor page that shows the correct figure.