LLM Cost Comparison Calculator
Compare monthly token costs across major LLM models — GPT-4o, Claude, Gemini, and more — for your specific workload.
Did this tool work for you?
How to use this calculator
- 1
Enter the average number of input tokens per request (your prompt + context).
- 2
Enter the average number of output tokens per request (model response length).
- 3
Enter your expected number of requests per month.
- 4
Review the monthly cost across all major LLM models to find the best fit.
Frequently asked questions
What formula and units does the comparison use?
For each model, monthly cost = requests × (input tokens × input rate + output tokens × output rate) ÷ 1,000, where the rates are dollars per 1,000 tokens. The built-in table below also shows the per-million equivalents, which are the rates most pricing pages quote.
Which models and prices are included?
A fixed list of five entries: GPT-4o, GPT-4o-mini, Claude Sonnet, Claude Haiku and Gemini 1.5 Pro, with the rates shown on this page. They are figures built into this tool. The list does not update itself, may not include newer or renamed models, and the Claude entries carry no version number. Check each provider’s pricing page.
How is the cheapest model chosen?
It is the lowest of the five monthly totals. On a tie the earlier entry in the list is named. If tokens or requests are zero, every total is $0.00 and the tool asks you to enter a workload instead of naming a model.
Can I add my own model or price?
No, the list is fixed. The AI Prompt Cost Calculator has a custom-pricing option that lets you enter your own input and output prices for one call.
Are the same token counts really valid for every model?
The tool applies the same input and output token counts to all five models, but each vendor uses its own tokenizer, so identical text can be a different length on each. Count tokens separately for each candidate if the difference matters.
Does the comparison say which model is better?
No. It compares price only. Quality, speed, context window and fit for your task are not measured, so test your own prompts on the models you shortlist.
What does it leave out?
Cached-input pricing, batch discounts, long-prompt tiers, image or tool-call charges, retries, free tiers and taxes. Every request is priced at flat input and output rates.
Why are the rates per 1,000 tokens here but per million elsewhere?
This calculator’s code stores rates per 1,000 tokens. Multiply by 1,000 to compare with per-million prices; for example 0.0025 per 1,000 equals $2.50 per million.
LLM cost comparison calculator: one workload priced across five built-in models
How the five totals are produced
Enter input tokens per request, output tokens per request and requests per month. For every model in the list, monthly cost = requests × (input tokens × input rate + output tokens × output rate) ÷ 1,000, with rates stored per 1,000 tokens. The page lists the five totals in a fixed order and names the lowest as the cheapest model for the workload. On a tie the earlier entry in the list wins, and with an empty workload the tool asks for numbers instead of naming a winner.
Rates stored in the code
These are figures built into this tool, not live quotes. The list is fixed at five entries and does not track new releases, so a newer or renamed model may be missing. The Claude entries carry no version number, which means the tool cannot tell you which release a rate belongs to. Check each provider’s pricing page before using any result in a budget.
| Model entry | Input per 1,000 | Output per 1,000 | Input per 1M | Output per 1M |
|---|---|---|---|---|
| GPT-4o | $0.0025 | $0.01 | $2.50 | $10 |
| GPT-4o-mini | $0.00015 | $0.0006 | $0.15 | $0.60 |
| Claude Sonnet | $0.003 | $0.015 | $3 | $15 |
| Claude Haiku | $0.00025 | $0.00125 | $0.25 | $1.25 |
| Gemini 1.5 Pro | $0.00125 | $0.005 | $1.25 | $5 |
Three workloads
Each row was produced by running the calculator. Values are US dollars per month.
| Workload | GPT-4o | GPT-4o-mini | Claude Sonnet | Claude Haiku | Gemini 1.5 Pro | Lowest |
|---|---|---|---|---|---|---|
| 1,000 in, 500 out, 10,000 requests (defaults) | 75.00 | 4.50 | 105.00 | 8.75 | 37.50 | GPT-4o-mini |
| 4,000 in, 200 out, 100,000 requests | 1,200.00 | 72.00 | 1,500.00 | 125.00 | 600.00 | GPT-4o-mini |
| 300 in, 1,500 out, 20,000 requests | 315.00 | 18.90 | 468.00 | 39.00 | 157.50 | GPT-4o-mini |
Reading the three workloads
The second workload is input-heavy with 4,000 tokens in and 200 out; the third is output-heavy with 300 in and 1,500 out. The order of the five totals is the same in all three, from GPT-4o-mini lowest to Claude Sonnet highest, which tells you that with these particular rate pairs, changing the input-to-output mix does not reorder the list. That could change if you compared models whose output rate is a different multiple of the input rate. The dollar gaps do move with the mix, so rerun the calculator with your own averages rather than relying on a ranking.
Limits when you compare
The same token counts are applied to every model, but vendors tokenize differently, so the same prompt may be a different length on each. Cached-input pricing, batch discounts, long-prompt tiers and tool or image charges are absent. Price is one input to a decision. The tool says nothing about quality, speed or context size for any model, so test your own prompts on the shortlisted candidates and compare the cost of errors as well as the cost of tokens.
Learn more from an authoritative source:
OpenAI Platform DocsEmbedding Cost Calculator
Calculate monthly and annual costs of generating text embeddings with OpenAI or Cohere models.
Fine-Tuning Cost Calculator
Estimate the training cost and ongoing inference cost of fine-tuning an OpenAI language model.
AI Image Generation Cost Calculator
Calculate monthly costs for AI image generation with DALL-E 3, Stable Diffusion, or Midjourney.
RAG Pipeline Cost Calculator
Estimate the monthly cost of a Retrieval-Augmented Generation (RAG) pipeline including indexing, embeddings, and LLM inference.
GPU Cloud Cost Calculator
Compare GPU rental costs across AWS, GCP, Azure, Lambda Labs, and Vast.ai for ML training and inference workloads.
AI Chatbot ROI Calculator
Calculate the return on investment of deploying an AI chatbot for customer support, including payback period and annual ROI.
Results are estimates for informational purposes only and do not constitute professional financial, medical, legal, or technical advice. Read full disclaimer →