Qwen API in India — Full Lineup Pricing

Alibaba's Qwen models rank among the best open-weight LLMs available today. On unoblox, access the full Qwen lineup — from the free 1.7B to the 235B and Max flagships — billed in rupees, no international card needed.

The Qwen Lineup (All Live)

Qwen spans free chat (Qwen3 1.7B), balanced reasoning (Qwen3 235B), open-weight vision (Qwen3.8-27B), and frontier reasoning (Qwen3 Max). Pick your tier:

Model	Use Case	Input	Output	Availability
Qwen3 1.7B	Fast chat, classification	₹0	₹0	Free tier
Qwen3 235B	Balanced reasoning + speed	₹9.07	₹55.44	Live
Qwen3.8-27B	Vision, open-weight, code	₹16.32	₹48.96	Live
Qwen3 Max	Frontier reasoning	₹120.95	₹604.77	Live

Why Choose Qwen on unoblox

• Open-weight: Qwen3.8-27B and Qwen3 235B are downloadable and fine-tunable on your own GPU.
• Vision built in: Qwen3.8-27B reads images, charts, and diagrams natively — no separate vision model or endpoint.
• India-first pricing: Qwen3 1.7B is free; the rest are priced well below most closed-source equivalents.
• One API key: switch between Qwen, GPT, Claude, and DeepSeek without rewriting a line of integration code.

Quick Start

1. Sign up → https://unoblox.ai/sign-in.
2. Copy your key → "API Keys" → select or create a key.
3. Call Qwen → use the matching model id in your OpenAI-compatible request.

Example (Python):

from openai import OpenAI
client = OpenAI(api_key="ub-gw-...", base_url="https://api.unoblox.ai/v1")
# Free tier: Qwen3 1.7B
response = client.chat.completions.create(
    model="qwen/qwen3-1.7b",
    messages=[{"role": "user", "content": "Summarize this article in three bullet points."}]
)

Qwen3 Max vs 235B vs Qwen3.8-27B — Picking a Model

Use Qwen3 Max for frontier reasoning and the hardest tasks. Use Qwen3 235B when you want strong general reasoning at a fraction of Max's price. Use Qwen3.8-27B when you need vision, want to fine-tune on your own data, or need a cheap workhorse for high-volume calls. For coding specifically, Qwen3.8-27B and Qwen3 235B both handle generation and review well — see /models for the current catalog and rates.

Frequently asked questions

Q: Can I download Qwen3.8-27B and run it locally? A: Yes. It's open-weight on Hugging Face — fine-tune it with your own data, or use unoblox for managed inference at ₹16.32 per million input tokens.

Q: What's the latency like? A: Streaming starts quickly and tokens arrive continuously; exact speed depends on prompt size, output length, and load. For latency-sensitive apps, start with Qwen3 1.7B or Qwen3.8-27B.

Q: Is Qwen3 1.7B good enough for production? A: It's surprisingly capable for chat, summarization, classification, and light reasoning. Move to 235B or Max for frontier reasoning tasks.

Q: Do Qwen models support function calling and tools? A: Yes — Qwen3 Max, Qwen3 235B, and Qwen3.8-27B all support the standard tools format.

Q: Can I use Qwen with my existing framework? A: Yes. LangChain, LlamaIndex, and the Vercel AI SDK all work — just point base_url to unoblox and use the Qwen model id.

Get started in rupees → https://unoblox.ai/sign-in