Claude API in India: Price in ₹ and How to Pay with UPI (2026)
Claude Sonnet 5 costs ₹207.22 per 1M input tokens and ₹1,036 per 1M output tokens on Velona (prices as of 4 October 2026). You can pay with UPI from ₹10, with no international credit card, and call Claude from Python or any OpenAI-compatible tool with one API key.
This guide covers the rupee price of every Claude model on Velona, what a typical request actually costs, exactly how the price is calculated, and how to make your first Claude API call from India in about five minutes.
Claude API prices in India (₹ per 1M tokens)
| Model | Model ID | Input ₹ / 1M tokens | Output ₹ / 1M tokens | Context window |
|---|---|---|---|---|
| Claude Haiku 4.5 | anthropic/claude-haiku-4.5 | ₹103.61 | ₹518.05 | 200K tokens |
| Claude Sonnet 5 | anthropic/claude-sonnet-5 | ₹207.22 | ₹1,036 | 1M tokens |
| Claude Opus 4.8 | anthropic/claude-opus-4.8 | ₹518.05 | ₹2,590 | 1M tokens |
| Claude Fable 5 | anthropic/claude-fable-5 | ₹1,036 | ₹5,180 | 1M tokens |
Prices follow the live exchange rate, so they move slightly. The Claude Sonnet 5 page and the other model pages always show the current figure.
What a Claude request actually costs in rupees
Per-million prices are hard to picture. Here is what common requests cost at today's rates:
| Request | Tokens (input / output) | Haiku 4.5 | Sonnet 5 | Opus 4.8 |
|---|---|---|---|---|
| A chat reply | 1,000 / 500 | ₹0.36 | ₹0.73 | ₹1.81 |
| Summarise a 10-page document | 6,000 / 600 | ₹0.93 | ₹1.86 | ₹4.66 |
| One step of a coding agent | 30,000 / 1,000 | ₹3.63 | ₹7.25 | ₹18.13 |
Two things stand out. Output tokens cost five times more than input tokens on every Claude model, so long answers are what drive your bill. And coding agents are expensive because they resend your code as input on every step. If you use Claude inside Cline, Cursor-style tools or your own agent, read our guide to the cheapest models for coding agents before you pick a default.
How the rupee price is calculated
Velona's price for every model follows one published formula:
Provider's USD price × live USD/INR rate × 1.005 (forex buffer) × 1.07 (platform fee)- The exchange rate is fetched live, so you pay close to the real rate.
- The 0.5% forex buffer covers rate movement between refreshes.
- The 7% platform fee covers infrastructure, support and payment processing.
- There is no subscription, no monthly minimum and no per-seat charge. You pay per token from a prepaid ₹ wallet.
To be straightforward about it: if you already have an international card that works with Anthropic and you're happy paying in dollars, going direct can cost less than paying the 7% fee. Velona is built for the many Indian developers who don't have that card, hit payment declines, or want to start with ₹10 instead of a dollar minimum, pay by UPI, and use Claude alongside 300+ other models through one key.
How to use the Claude API from India in 5 minutes
Step 1: Create an account and top up with UPI
Sign up at velona.in, open the wallet and add money with UPI, RuPay, a domestic card or net banking. The minimum top-up is ₹10, which covers about 13 Sonnet 5 chat replies or about 27 on Haiku 4.5.
Step 2: Create an API key
Go to API keys and create a key for your project. You can give each key its own budget so one project can't drain your wallet.
Step 3: Call Claude from Python
Velona is OpenAI-compatible, so you use the standard OpenAI Python SDK and only change the base URL and the model ID.
pip install openai python-dotenvPut your key in a .env file and keep that file out of Git:
VELONA_API_KEY=your_key_hereimport os
from dotenv import load_dotenv
from openai import OpenAI
load_dotenv()
client = OpenAI(
api_key=os.environ["VELONA_API_KEY"],
base_url="https://velona.in/v1",
)
response = client.chat.completions.create(
model="anthropic/claude-sonnet-5",
messages=[
{"role": "user", "content": "Explain recursion with a short Python example."}
],
max_tokens=500,
)
print(response.choices[0].message.content)
print(response.usage)Setting max_tokens caps the most expensive part of the request, and response.usage shows exactly how many tokens you were billed for.
Step 4: Or test it with curl
curl https://velona.in/v1/chat/completions \
-H "Authorization: Bearer $VELONA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "anthropic/claude-haiku-4.5",
"messages": [{"role": "user", "content": "Say hello in Hindi and English."}],
"max_tokens": 100
}'Every request shows up in your usage dashboard with its cost in rupees.
Which Claude model should you use?
- Claude Haiku 4.5 for chatbots, classification, extraction and anything where speed and cost matter more than depth. It is half the price of Sonnet 5.
- Claude Sonnet 5 as the default for most coding, writing and analysis. It has a 1M-token context window, so it can take in large codebases and long documents.
- Claude Opus 4.8 for the hardest multi-step reasoning and long agent runs, at 2.5 times the price of Sonnet 5.
- Claude Fable 5 only when you need the strongest coding result available. It has the highest coding score on our coding leaderboard, but it costs five times as much as Sonnet 5.
A pattern that works well: send simple requests to Haiku 4.5, and only route the hard ones to Sonnet 5 or Opus 4.8. Because every model sits behind the same API, switching is a one-line change to the model value.
How to keep your Claude bill low
- Set
max_tokenson every request. Output costs five times more than input. - Send only the context the model needs, not the whole conversation or file every time.
- Use Haiku 4.5 for first drafts, routing and simple questions.
- Give each API key its own budget, especially for experiments and student projects.
- Compare models on your own prompt with the model catalogue before you commit to one.
Frequently asked questions
Can I use the Claude API in India?
Yes. Through Velona you can call Claude Haiku, Sonnet, Opus and Fable models from India, pay in rupees with UPI, and use the same OpenAI-compatible code you would use anywhere else.
How much does 1M tokens of Claude cost in rupees?
On Velona today, 1M input tokens cost ₹103.61 on Haiku 4.5, ₹207.22 on Sonnet 5, ₹518.05 on Opus 4.8 and ₹1,036 on Fable 5. Output tokens cost five times the input price on each model.
Do I need an international credit card for the Claude API?
Not with Velona. You top up a rupee wallet with UPI, RuPay, a domestic debit or credit card, or net banking.
What is the minimum amount to start?
₹10. There is no subscription and no monthly minimum, and you only pay for the tokens you use.
Can I use Claude in coding tools like Cline or OpenCode?
Yes. Any tool that accepts an OpenAI-compatible provider works. Set the base URL to https://velona.in/v1, paste your Velona key, and use a model ID such as anthropic/claude-sonnet-5.
Start using Claude in rupees
Create a free account, add ₹10 with UPI, and send your first Claude request in a few minutes. The same key also works with DeepSeek, GPT, Gemini, Grok and 300+ other models, all priced in rupees.