LLM API Pricing & Token Cost Calculator
Compare monthly inference costs across OpenAI (GPT-4o, o1), Anthropic (Claude 3.5 Sonnet, Haiku), Google (Gemini 2.0 Flash, 1.5 Pro), and DeepSeek (V3, R1). Calculate exact compute budgets in USD ($) and INR (₹).
DeepSeek-V3
Claude 3.5 Sonnet
Gemini 2.0 Flash
Frequently Asked Questions: LLM API Pricing
Which LLM API is the cheapest in 2026?
DeepSeek-V3 ($0.14 input / $0.28 output per 1M tokens) and Google Gemini 2.0 Flash ($0.10 input / $0.40 output per 1M tokens) are currently the most affordable flagship-tier models available, delivering performance comparable to GPT-4o at less than 10% of the cost.
What is the difference between Input (Prompt) and Output (Completion) tokens?
Input tokens are the words, instructions, and context you send to the model (including system prompts, documents, and past conversation history). Output tokens are the generated response words that the model computes. Output tokens generally cost 3x to 4x more than input tokens because generating each subsequent token requires sequential auto-regressive GPU processing.
How can I reduce my monthly AI API costs by 50% or more?
Three proven strategies: (1) Prompt Caching: Keep system prompts static to earn 50-80% discounts on cache reads. (2) Model Routing: Use ultra-fast models like GPT-4o mini or DeepSeek-V3 for simple intent classification, and only route complex coding or reasoning prompts to Claude 3.5 Sonnet or o1. (3) Batch API: For async background jobs, OpenAI and Anthropic offer a flat 50% discount for results delivered within 24 hours.
