AI API Pricing Reference
Compare pricing, availability and capabilities across leading AI models and providers. Updated weekly.
Calculate API Costs
Monthly Cost Estimate
Based on published API pricing. Actual costs may vary.
Recent Changes
Answer 5 questions about your use case, volume, and budget to get a personalized model recommendation.
128 models compared, key pricing trends, and how to cut costs by 40-96%.
Answer 3 questions and get the cheapest model for your use case with instant cost estimates.
7 proven strategies to cut costs by 40-98%, with real pricing data from 128 models.
Visual timeline of every model's lifecycle — releases, updates, deprecations, replacements.
Track all 128 models across 17 providers — status, pricing, context windows, migration paths.
Tools
Data verified · models tracked · providers
FAQ
Our calculator uses published API pricing from each provider, updated regularly. Results are estimates based on token counts — actual costs may vary slightly depending on your specific usage patterns, but they're typically within 5-10% of your real bill.
APIpulse supports 128 models across 17 providers: OpenAI, Anthropic, Google, DeepSeek, Mistral, Cohere, Meta (Muse Wonder), Meta Llama (via Together.ai), Moonshot, xAI, Xiaomi MiMo, AI21, Qwen, Z.ai, plus InclusionAI, Inception and IBM Granite. We cover both premium models (GPT-6 Astra, Claude Opus 5.5, Gemini 4 Argon) and budget options (GPT-6 Luna, Claude Haiku 5.5, Gemini 2.5 Flash-Lite, DeepSeek V4.1 Flash).
The biggest savings come from using the right model for each task — route simple requests to budget models like GPT-6 Luna and reserve premium models for complex reasoning. Other strategies include prompt optimization, response caching, batch processing, and setting token limits. Our calculator helps you model these scenarios.
Yes — the cost calculator, provider comparison table, and all tools and blog content are completely free. No signup, no credit card required.
We track published pricing pages from each provider and update the dataset weekly. Each model entry shows a "verified" date so you can see how current the data is.