Compare LLM API costs across 456+ models — GPT-4o, Claude 3.5, Gemini, Llama, Mistral, DeepSeek and more. Click any model to open its dedicated pricing page. Updated Oct 29, 2026.
per request
per request
daily volume
LLM API Pricing — 456+ Models Compared
This calculator covers 456 LLM models from all major providers. Pricing data is sourced from OpenRouter and updated each time this site is deployed. Click any model row for a detailed cost breakdown and usage estimates.
How LLM Pricing Works
Every major LLM API charges per token — a token is roughly 0.75 words. Pricing has two components:
Input (prompt) tokens: Everything you send to the model — system prompt, conversation history, user message.
Output (completion) tokens: What the model generates. Usually 2–4x more expensive per token than input.
The table shows 50 models per page sorted by cost. Use Prev/Next buttons or the page number links to browse all 456 models. Use the search box to filter by name or provider first. Click any model row to open its dedicated pricing page with full cost breakdown and usage estimates.
Free vs Paid LLMs
Several models show “Free” for input and output pricing. These are free inference endpoints on OpenRouter with rate limits (typically 20 req/min or 200/day). Ideal for prototyping; for production, paid models with SLAs are recommended.
Frequently Asked Questions
Pricing is fetched from the OpenRouter public API at build time and embedded directly in this page. It covers 456+ models. Last updated Oct 29, 2026.
A token is roughly 0.75 words or 4 characters. 1,000 tokens ≈ 750 words. APIs charge separately for input (prompt) tokens and output (completion) tokens.
Click any row in the table to open a detail popup showing input/output rates, context window size, per-request cost, and estimates for common usage volumes.
Use the Prev/Next buttons or page number links below the table. Each page shows 50 models. Search first to narrow results, then paginate through the filtered list.
The context window is the maximum number of tokens (input + output combined) the model can process in one request. A 128K context window can handle roughly 96,000 words of combined prompt and response.
OpenRouter is an API gateway providing access to 200+ LLMs under one API key. It aggregates models from OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek, and many others.