LLM API Pricing Comparison

Every major model. One sortable table.

Interactive comparison of API pricing across all major LLM providers. Sort by input cost, output cost, or context window size. Filter by provider or model category. All prices shown per 1 million tokens in USD.

Providers

Category

Current LLM Pricing Landscape

LLM API pricing spans a wide range. Budget models like GPT-5 nano ($0.05 per million input tokens) and Gemini 2.0 Flash Lite ($0.08) serve high-volume, low-complexity tasks. Mid-tier models like Claude Sonnet 4.6 ($3.00) and Gemini 3 Flash ($0.50) balance capability with cost. Flagship models like GPT-5.4 ($2.00), Claude Opus 4.6 ($5.00), and Gemini 3.1 Pro ($2.00) offer the strongest reasoning at premium prices. Pricing has converged significantly over the past year as competition intensifies.

How to Read This Table

All prices are shown per 1 million tokens in USD. Input tokens are what you send to the model. Output tokens are what the model generates. Most providers also offer cached input pricing (discounted rates for repeated prompt prefixes) and batch pricing (discounted rates for asynchronous bulk processing). Click any column header to sort. Use the filters above the table to narrow your view by provider or model tier.

Frequently Asked Questions

What is the cheapest LLM API in 2026?
the cheapest LLM APIs by input token price are GPT-5 nano at $0.05 per million tokens, Gemini 2.0 Flash Lite at $0.08, and Mistral Small at $0.10. For output tokens, GPT-5 nano ($0.40/M) and Gemini 2.0 Flash Lite ($0.30/M) are the most affordable.
Why are output tokens more expensive than input tokens?
Output tokens require more computation than input tokens because the model must generate each token sequentially, running inference for every token produced. Input tokens are processed in parallel. This computational difference is reflected in pricing, with output tokens typically costing 2-5x more than input tokens.
What does 'per 1M tokens' mean in pricing?
LLM API pricing is quoted per 1 million tokens. To calculate your cost for a specific request, divide your token count by 1,000,000 and multiply by the per-million price. For example, processing 10,000 input tokens on GPT-5.4 at $2.00/1M costs: (10,000 / 1,000,000) × $2.00 = $0.02.
How does GPT-5.4 pricing compare to Claude Opus 4.6?
GPT-5.4 costs $2.00 per million input tokens and $8.00 per million output tokens. Claude Opus 4.6 costs $5.00 input and $25.00 output per million tokens. GPT-5.4 is significantly cheaper per token, though the models have different strengths. Use our Model Selector tool to determine which is the better fit for your specific use case.

All prices in USD per 1 million tokens. Prices sourced from official provider pricing pages. Last verified on the date shown above.