Data sources and estimate limits

Updated November 9, 2026.

Sources and scope

Each published model links to an official provider price page and includes a region or routing condition. Prices are in USD for direct API access. Provider subscriptions and reseller plans are different products.

Included

• Input/output Token costs and an estimated cache-read share.
• DeepSeek peak-time weighting across input, cache reads and output.
• Verified per-model batch input/output rates. Models without verified batch prices are excluded from batch comparisons.
• OpenAI long-input multipliers and Qwen’s listed input tiers, based on average input length. Real request lengths may vary.

Excluded

Cache-write premiums, cache storage, tool calls, taxes, free credits, negotiated discounts, gateway fees and model-specific cache eligibility. Batch mode does not combine caching discounts. Custom rates are user assumptions, not provider quotes.

Price changes

Current temporary prices are labeled in the rate conditions. Verify the price for your intended usage period. Expired dated rates are excluded unless you supply custom prices. Shared links contain inputs and a price version; when the price version changes, a warning appears and current rates are used.

Reporting an issue

Contact information. Confirm prices on the provider’s own page before purchasing.