Methodology
What the prices are
All prices are standard API list prices in US dollars per one million tokens, with input and output shown separately. Where a provider changes price above a prompt-size threshold or on a set date, we note it next to the price.
Where the Anumbers come from
We start from the providers' own pricing pages and cross-check them against launch announcements and independent price trackers. Official pages come first. When sources disagree and we cannot settle it, we mark the model as unconfirmed rather than pick a number.
Official pricing pages: Anthropic · OpenAI · Google Gemini · xAI · Mistral · DeepSeek · Amazon Bedrock · Cohere.
What is not included
- Batch and cached-input discounts, unless a page says otherwise
- Regional or priority-tier surcharges
- Taxes such as VAT, and currency conversion fees
- Tool, search and fine-tuning fees
- Provider-specific enterprise or volume contracts
How often we update
We review the price table at least once a month and when a provider announces a change, for example a scheduled price increase. Each page shows the date its prices were last checked. Prices can change between reviews, so confirm on the provider's page before you commit.
How the cost figures are calculated
Cost equals input tokens times the input price plus output tokens times the output price, divided by one million. Worked examples use the stated usage, such as 1,000 requests a day with 1,500 input and 400 output tokens over 30 days. Read the full formula in how to calculate LLM API cost.
Independence and disclosure
Providers do not pay for placement. Some links are affiliate links and may earn a commission. Drafts are prepared with AI assistance and reviewed by the publisher before publication.
Corrections
If you spot an error, email grisvardcedric@gmail.com. We correct confirmed errors promptly.