OpenAI API in India: GPT-5 Price in ₹ and How to Pay with UPI (2026)
You can use the OpenAI API from India without an international credit card. On Velona, GPT-5 Mini costs ₹25.81 per 1M input tokens and ₹206.46 per 1M output tokens, GPT-5.1 costs ₹129.04 and ₹1,032, and you top up a rupee wallet with UPI starting at ₹10 (prices as of 6 October 2026). Your existing OpenAI SDK code works unchanged once you point it at https://velona.in/v1.
This guide covers why Indian cards are so often declined on OpenAI, the rupee price of every GPT model, what a real request costs, and how to send your first GPT request from India in about five minutes.
Quick answer
- Can you use the OpenAI API in India? Yes. You can pay OpenAI directly with a card that works for international recurring payments, or use an INR gateway like Velona and pay with UPI.
- Cheapest GPT model: GPT-5 Nano at ₹5.16 per 1M input tokens, or the open-weight GPT-OSS 120B at ₹3.82.
- Best value default: GPT-5 Mini. A typical chat reply costs about ₹0.13.
- Minimum spend: ₹10. No subscription and no monthly minimum.
Why the OpenAI API keeps declining Indian cards
OpenAI bills API usage in US dollars and accepts cards only. Many Indian developers see "Your card has been declined" even when the card works on other foreign websites. The OpenAI developer forum has long threads on this, and the common explanations are RBI rules on recurring payments and e-mandates, 3D Secure failures at checkout, and international transactions being switched off on the card.
Some people get through by trying a different bank's credit card or enabling international usage. Many never do, and debit cards, RuPay and UPI are not accepted at all. That is the gap Velona fills: you pay in rupees through Razorpay with UPI, RuPay, a domestic card or net banking, and Velona handles the dollar side.
OpenAI API prices in India (₹ per 1M tokens)
| Model | Model ID | Input ₹ / 1M tokens | Output ₹ / 1M tokens | Best for |
|---|---|---|---|---|
| GPT-OSS 120B | openai/gpt-oss-120b | ₹3.82 | ₹17.55 | Lowest cost, open weights |
| GPT-5 Nano | openai/gpt-5-nano | ₹5.16 | ₹41.29 | Classification, extraction, routing |
| GPT-4o Mini | openai/gpt-4o-mini | ₹15.48 | ₹61.94 | Existing apps built on 4o Mini |
| GPT-5 Mini | openai/gpt-5-mini | ₹25.81 | ₹206.46 | Everyday default for chat and apps |
| o4-mini | openai/o4-mini | ₹113.55 | ₹454.21 | Cheaper reasoning |
| GPT-5 | openai/gpt-5 | ₹129.04 | ₹1,032 | Hard reasoning and coding |
| GPT-5.1 | openai/gpt-5.1 | ₹129.04 | ₹1,032 | Latest flagship at the same price as GPT-5 |
| o3 | openai/o3 | ₹206.46 | ₹825.85 | Deep multi-step reasoning |
| GPT-4o | openai/gpt-4o | ₹258.08 | ₹1,032 | Legacy apps only |
Prices follow the live exchange rate, so they move slightly every day. Each model page shows the current figure and a monthly cost calculator. One thing worth noticing: GPT-4o now costs more per input token than GPT-5.1, so if you are still on 4o, moving to GPT-5.1 or GPT-5 Mini is usually both cheaper and better.
What a GPT request actually costs in rupees
| Request | Tokens (input / output) | GPT-5 Nano | GPT-5 Mini | GPT-5.1 |
|---|---|---|---|---|
| A chat reply | 1,000 / 500 | ₹0.03 | ₹0.13 | ₹0.65 |
| Summarise a 10-page document | 6,000 / 600 | ₹0.06 | ₹0.28 | ₹1.39 |
| One step of a coding agent | 30,000 / 1,000 | ₹0.20 | ₹0.98 | ₹4.90 |
So ₹10 buys roughly 75 GPT-5 Mini chat replies or more than 350 on GPT-5 Nano. On GPT-5 family models, output tokens cost eight times more than input tokens, so long answers drive your bill far more than long prompts do.
How the rupee price is calculated
Every model on Velona uses one published formula:
OpenAI's USD price x live USD/INR rate x 1.005 (forex buffer) x 1.07 (platform fee)- The exchange rate refreshes every 10 minutes.
- The 0.5% buffer protects against rate movement between refreshes.
- The 7% fee covers infrastructure, payment processing and support.
- No subscription, no monthly minimum, no per-seat fee, and no foreign currency markup on your bank statement.
To be honest about the trade-off: if your card already works with OpenAI and you spend a lot each month, paying OpenAI directly in dollars can cost less than the 7% fee, even after your bank's forex charges. Velona is the better fit when your card is declined, you want to pay by UPI, you want to start with ₹10 instead of a dollar top-up, or you want GPT, Claude, Gemini, DeepSeek and 300+ other models behind one key.
How to use the OpenAI API from India in 5 minutes
Step 1: Create an account and top up with UPI
Sign up at velona.in with Google, GitHub or email. Open the wallet and add money with UPI, RuPay, a domestic card or net banking, from ₹10.
Step 2: Create an API key
Go to API keys and create a key. You can give each key its own spending limit so a test script can never drain your wallet.
Step 3: Call GPT from Python with the official OpenAI SDK
pip install openai python-dotenvSave 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="openai/gpt-5-mini",
messages=[
{"role": "user", "content": "Explain list comprehensions in Python with one example."}
],
max_tokens=400,
)
print(response.choices[0].message.content)
print(response.usage)The only two changes from a normal OpenAI script are base_url and the openai/ prefix on the model ID. Everything else, including streaming and the usage object, works the same.
Step 4: Or test with curl
curl https://velona.in/v1/chat/completions \
-H "Authorization: Bearer $VELONA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-5-nano",
"messages": [{"role": "user", "content": "Translate good morning into Hindi, Bengali and Tamil."}],
"max_tokens": 100
}'Every request appears in your usage dashboard with its exact cost in rupees.
Using JavaScript or Node.js?
The same two changes apply to the official openai npm package. Our JavaScript guide walks through it.
Which GPT model should you use?
- GPT-5 Nano for classification, tagging, extraction and routing, where you send many small requests.
- GPT-5 Mini as the default for chatbots, student projects, content tools and most apps.
- GPT-5.1 when quality matters most: complex coding, analysis and long documents.
- o4-mini or o3 for step-by-step reasoning such as maths and planning.
- GPT-OSS 120B when cost matters more than anything else and you want an open-weight model.
Cheaper alternatives to GPT for the same job
Because one Velona key works with every model, you can compare GPT against other providers by changing a single line. A few that Indian developers often switch to:
| Model | Input ₹ / 1M | Output ₹ / 1M | Why people pick it |
|---|---|---|---|
| DeepSeek V4 Flash | ₹4.32 | ₹8.64 | Very low output price, strong for coding agents |
| Gemini 2.5 Flash Lite | ₹10.32 | ₹41.29 | Fast, cheap, long context |
| Claude Haiku 4.5 | ₹103.23 | ₹516.15 | Reliable instruction following |
See the full list on cheapest AI models, or compare quality on the benchmark page.
Frequently asked questions
Is the OpenAI API available in India?
Yes. India is a supported country. The usual problem is payment, not access. If OpenAI declines your card, you can use GPT models through Velona and pay in rupees with UPI.
How much does the OpenAI API cost in rupees?
On Velona as of 6 October 2026, 1M input tokens cost ₹5.16 on GPT-5 Nano, ₹25.81 on GPT-5 Mini and ₹129.04 on GPT-5 and GPT-5.1. Output tokens cost eight times the input price on these models.
Can I pay for the OpenAI API with UPI?
Not directly, since OpenAI accepts only cards for API credits. Through Velona you top up a rupee wallet with UPI, RuPay, a domestic debit or credit card, or net banking, and use that balance for GPT requests.
Why does OpenAI say my card has been declined?
Common causes are international transactions being disabled on the card, RBI e-mandate rules for recurring charges, and 3D Secure failures during checkout. Enabling international usage or trying a different bank's credit card sometimes helps. If it doesn't, an INR gateway avoids the problem entirely.
Will my existing OpenAI code work with Velona?
Yes. Velona is OpenAI-compatible. Change the base URL to https://velona.in/v1, use your Velona key, and add the openai/ prefix to the model name. It also works in tools that accept a custom OpenAI provider, such as Cline, Continue and OpenCode.
What is the minimum amount to start?
₹10. There is no subscription or monthly minimum, and you only pay for tokens you use.
Start using GPT in rupees
Create a free account, add ₹10 with UPI, and send your first GPT-5 request in a few minutes. The same key also gives you Claude, Gemini, DeepSeek, Grok, Llama and 300+ other models, all priced in rupees.