GPT OSS 120B Turbo API

Pricing

₹20/1M input · ₹90/1M output. Pay-as-you-go micro-metered billing. UPI and cards accepted with zero minimum commitment.

OpenAI-Compatible Integration

Drop into existing OpenAI SDK code by changing the base URL to https://api.luminoai.co.in/v1. The request shape below is the reasoning pattern.

curl https://api.luminoai.co.in/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $LUMINO_API_KEY" \
  -d '{
    "model": "gpt-oss-120b-turbo",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Work through this problem step by step and give the final answer."}
    ]
  }'

Good fit for reasoning

• Problems that need multi-step reasoning before the answer
• Maths, logic, and analysis where a wrong first attempt is costly
• Planning tasks where the model must weigh several options
• Evaluation and review work that needs justification, not just output

Model Specifications

Why reasoning models bill more than they appear to

Reasoning models generate hidden reasoning tokens before the visible answer, so the output count you see is not the output count you pay for. Long chains also mean higher latency. Budget for output tokens two to three times the visible answer length.

GPT OSS 120B Turbo API FAQ

Why is GPT OSS 120B Turbo slower than a chat model?

It produces reasoning tokens before the final answer. That extra generation is billed and adds latency.

Can I see the reasoning?

The visible answer is what the API returns by default. Check the model catalog entry for reasoning-trace support on this specific model.

When should I not use a reasoning model?

For short factual replies, classification, and formatting tasks. A cheaper chat tier returns the same result for a fraction of the output cost.