Llama 4 Maverick API in India

Llama 4 Maverick is Meta's larger, higher-reasoning model in the Llama 4 family, reachable through unoblox's OpenAI-compatible gateway in India. At ₹20.16 per million input tokens and ₹80.64 per million output tokens, it delivers near-proprietary-grade quality at a fraction of GPT-5 or Claude Opus pricing, with no vendor lock-in.

Pricing and technical profile

Per 1M tokens:

• Input: ₹20.16
• Output: ₹80.64
• Weights: open-source, available for local deployment alongside the managed API

Open weights mean unoblox's price is purely for inference — no separate licensing fee on top.

When Maverick is the right choice

• Complex reasoning on a budget — roughly 25x cheaper than Claude Opus on input, with quality that's competitive for most tasks.
• Multilingual work — strong performance on code, math, Hindi, and other Indian languages.
• Domain-specific fine-tuning — since it's open-source, you can adapt Maverick to your vertical without licensing friction.
• Long-context RAG — check /models for the current context limit; it's generous enough for document-heavy retrieval pipelines.

Two-minute integration

Base URL: https://api.unoblox.ai/v1
Model: meta/llama-4-maverick
API Key: ub-gw-[your-key]

Streaming and tool-calling work as expected — no code rewrites needed beyond the base URL and model string.

Maverick vs Scout

Factor	Scout	Maverick
Input (₹/1M)	₹10.08	₹20.16
Output (₹/1M)	₹30.24	₹80.64
Reasoning depth	Good	Higher
Best for	High-volume, simpler tasks	Complex reasoning, higher quality

Why Maverick on unoblox

Open-source freedom — no separate OpenAI or Anthropic account overhead; start today and scale without licensing hassle.

Transparent pricing — every rupee cost is visible upfront, with nothing to reconcile against a foreign statement.

GST invoice — monthly billing that fits standard Indian company accounting, input-tax credit claimable on full spend.

Frequently asked questions

How does Maverick compare to GPT-4o on cost? Maverick is about 12.5x cheaper than GPT-4o on both input and output (₹20.16 vs ₹252 input; ₹80.64 vs ₹1008 output) — GPT-4o still edges ahead on native vision, but for text reasoning Maverick is a serious budget option.

How does Maverick compare to Claude Opus? Roughly 25x cheaper on input and 31x cheaper on output — Opus has an edge on the hardest reasoning tasks, but Maverick covers most production workloads for a fraction of the price.

Can I use Maverick for production chat? Yes — it's responsive enough for chat interfaces, customer support, and agent systems.

Is the open-source license free for commercial use? Yes — Llama's community license permits commercial products without per-seat royalties; check Meta's current license terms for the specifics that apply to your use case.

What's the maximum context I can send? Check /models for the current limit on this model card; typical workflows use well under the maximum.

Does Maverick support vision? This version is text-only — for multimodal input, use GPT-4o or a Claude model on the same key.

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