LLM Token & Cost Calculator
Estimate API costs for the most popular AI models. Compare pricing, calculate monthly spend, and find the most cost-effective model for your use case.
Advertisement
Advertisement
2025 LLM Pricing Comparison
LLM API prices have dropped dramatically — approximately 80% across the industry from early 2025 to early 2026. This makes building AI-powered applications more affordable than ever. Here's a quick reference for the most popular models:
| Model | Input ($/1M) | Output ($/1M) | Context |
|---|
How to estimate your token usage
A helpful rule of thumb: 1 token ≈ 4 characters ≈ 0.75 words in English. So 1,000 tokens is roughly 750 words or about 1.5 pages of standard text.
- A typical chatbot message: 100–500 input tokens, 100–300 output tokens
- Summarizing a long document: 2,000–8,000 input tokens, 300–800 output tokens
- A RAG query with retrieved context: 2,000–6,000 input tokens
- An AI agent loop: can accumulate 10,000+ tokens over multiple steps
Frequently Asked Questions
What is the cheapest LLM API in 2025?
Gemini 1.5 Flash and GPT-4o Mini are among the most affordable at under $0.15/1M input tokens. For open-source models hosted via API, providers like Together AI and Groq offer Llama models at very competitive rates.
Why do input and output tokens cost different amounts?
Generating (output) tokens requires more computation than reading (input) tokens. The model must generate each output token sequentially, making it more resource-intensive — hence output costs 3–5x more per token on most models.
Does context window size affect cost?
Yes — every token in the context window counts as an input token for each request. If you maintain a long conversation history, all previous messages are resent each time, multiplying your input token usage.