What will an AI API cost me per month?

Enter how much your app or script will use the API. See the monthly bill in rupees for every major model, at the providers' published prices.

1,000 English words ≈ 1,300 tokens

Every model at your usage, cheapest first

Or a flat monthly subscription?

For your own use in the provider's app (chat, writing, coding). Usage limits apply, and a subscription cannot power your own app or scripts: that needs the API.

PlanPer monthWhat you get
Claude Pro₹2,217 (US$20 + 18% GST)More usage than Free, every 5 hours and every week ($17/month if paid yearly)
Claude Max₹11,084 (US$100 + 18% GST)5x Pro usage every 5 hours and every week
Claude Ultra₹22,167 (US$200 + 18% GST)No 5-hour limit; 10x Pro usage every week
Google AI Plus₹399Gemini 4 Hydrogen; 2x the usage of the free plan
Google AI Pro₹1,950Gemini 4 Argon and Hydrogen; 4x the usage of the free plan
Google AI Ultra₹6,500Gemini 4 Carbon; 5x AI Pro usage (₹19,500/month tier: 20x)
ChatGPT GoSee today's priceMore messages with tools; may include ads
ChatGPT PlusSee today's priceExpanded messages and uploads
ChatGPT ProSee today's priceChoice of 3 usage tiers

Prices from claude.com/pricing (US$, tax extra; converted at ₹93.93 and 18% GST added as an estimate) andgemini.google/subscriptions (₹, as shown for India), last checked 16 Nov 2026. ChatGPT plan prices could not be confirmed automatically; check chatgpt.com/pricing.

Questions people ask

What is a token?

AI APIs charge by tokens, small pieces of text. In English, 1,000 words is roughly 1,300 tokens. Hindi, Bengali and other Indian languages need more tokens for the same meaning (in our tests about 1.3 to 1.9 times with a current tokenizer, more with older ones), so they cost more. Try our token counter to see the difference. Input tokens are what you send (your prompt, documents, chat history); output tokens are what the model writes back, and they cost several times more.

Why are prices in dollars?

All these providers bill in US dollars. We convert at the day's reference rate shown under the results. If you pay by Indian card, your bank adds a foreign currency markup, and the invoice may include 18% GST: add both in "extra on the bill" to see your real cost.

What does "cached input" mean?

If many requests start with the same text (a long system prompt, a document, chat history), providers can cache it and charge 75 to 99% less for those tokens on later requests. Set the share of each request that repeats. Some providers also charge to write or store the cache, which this estimate leaves out.

How can I cut the bill?

Pick the smallest model that does the job well, keep prompts short, cache repeated context, and limit how long answers can be. Most providers also offer a batch mode at about half price for work that can wait a few hours.

Is a Pro or Max subscription cheaper than the API?

For your own chatting, writing and coding in the provider's app, often yes: a subscription is a flat monthly price with usage limits, while the API charges per token. But a subscription cannot power your own app, website or scripts; that needs the API. If your API estimate is above a plan's price and the use is just you, the plan is usually the better deal.

Would running a model on my own GPU be cheaper?

For heavy, steady use it can be. Open models such as Qwen or Gemma run free on your own graphics card apart from electricity. Use our "Which AI models can my GPU run?" and "Rent a cloud GPU or buy one?" tools to compare.

Spotted a wrong number? Tell us and we'll fix it.