Skip to content
Qwen API · India

Qwen API, India — billed in rupees.

Alibaba's Qwen3 family — Max, 235B-A22B, the open-weight 3.8-27B, and free Qwen3-1.7B — through one OpenAI-compatible endpoint. Rupee pricing, a GST invoice, and up to 262K context.

OpenAI-compatible · ₹-native billing · Free tier · Open weights

Available models

Both Qwen lines, in ₹

From a free small model to the flagship — one key covers them all. Prices are per million tokens; the platform fee and GST are added at billing.

Qwen3 1.7B
qwen/qwen3-1.7b
41K ctx

Input / 1M

Free

Output / 1M

Free

Text · reasoning · tools. A genuinely free model for prototyping and evals.

Qwen3 235B A22B
qwen/qwen3-235b-a22b-instruct-2507
262K ctx

Input / 1M

₹9.07/M

Output / 1M

₹55.42/M

Text · tools. High-volume workhorse at a very low input rate.

Qwen3.8-27B
qwen/qwen3.8-27b
262K ctx

Input / 1M

Free

Output / 1M

Free

Text · image · reasoning · tools. Open weights — run it on unoblox now, self-host later.

Qwen3 Max
qwen/qwen3-max
256K ctx

Input / 1M

₹120.93/M

Output / 1M

₹604.63/M

Text · reasoning · tools. Alibaba's flagship hosted model, for the hardest work.

Indicative pricing — the live rate for each model, and the platform fee, is on its model page and in the pricing docs.

Why here

Cheap at the bottom, open in the middle

  • Start free — Qwen3-1.7B costs nothing — build and evaluate before you fund a wallet.
  • Open weights — Qwen3.8-27B ships open — run it here now, self-host later.
  • Vision & tools — Qwen3.8-27B adds image input; the line supports tool calls and JSON.
  • ₹ GST invoice — Billed in rupees from an Indian entity, ITC where registered.
curl
curl https://api.unoblox.ai/v1/chat/completions \
  -H "Authorization: Bearer $UNOBLOX_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-235b-a22b-instruct-2507",
    "messages": [{ "role": "user", "content": "Classify this ticket and reply in Hindi." }]
  }'
Route smart

Mix Qwen with everything else

Route bulk work to a cheap Qwen model and reserve frontier models for the hard requests — one key, one rupee bill.

FAQ

Qwen in India, answered

Which Qwen models are available in India?

Qwen3-1.7B (free), Qwen3-235B-A22B, the open-weight Qwen3.8-27B (with vision) and the flagship Qwen3-Max — all via the OpenAI-compatible endpoint.

How much does Qwen cost in rupees?

Qwen3 1.7B is free for both input and output; Qwen3 235B A22B is ₹9.07 per 1M input and ₹55.42 per 1M output; Qwen3.8-27B is free for both input and output; Qwen3 Max is ₹120.93 per 1M input and ₹604.63 per 1M output. Indicative — live rates are on each model page.

Which Qwen model should I start with?

Start with the free Qwen3-1.7B or the low-cost Qwen3-235B-A22B for high-volume work; reach for Qwen3-Max when you need the flagship, or Qwen3.8-27B if you want open weights and vision.

Can I self-host Qwen later?

Qwen3.8-27B ships as open weights, so you can move it in-house later; hosted-only models like Qwen3-Max stay on the API. One key spans both while you decide.

Do I need to change my code?

No — set the base URL to https://api.unoblox.ai/v1 and the model id to a qwen/… slug. The OpenAI SDK, streaming, tool calls and JSON mode all work unchanged.

How is it billed, and where does it run?

In rupees on a GST invoice from an Indian entity, with ITC where registered. Inference runs on the provider's infrastructure; open models may run on unoblox GPUs in India. No international card is required.

Call Qwen in rupees

Create a key and run the whole Qwen3 line — free, open-weight and flagship — through one OpenAI-compatible endpoint, priced in ₹.