Models

GPT-oss-120b vs Sonar Reasoning Pro

API price, context window, capabilities and live speed, side by side.

GPT-oss-120b is about 13× cheaper than Sonar Reasoning Pro at a typical 3:1 input-to-output token mix. GPT-oss-120b has the larger context window (131K vs 128K tokens). Only Sonar Reasoning Pro supports image input; GPT-oss-120b does not. Both run on the same UltraGPT API key, so you can switch between them by changing the model field - or list both in models: [...] for automatic fallback.

GPT-oss-120bSonar Reasoning Pro
LabOpenAIOther
Model idgpt-oss-120bsonar-reasoning-pro
Input / 1M tokens$0.15$2
Cached input / 1M$0.075$2
Output / 1M tokens$0.6$8
Context window131K128K
Inputtexttext, image
Outputtexttext
Tool calling——
Vision (image input)—✓
Reasoning——

Cost of 1,000 requests

Estimated from the list prices above; your real cost is reported per request in usage.cost.

Chatbot1.5K tokens in, 400 outGPT-oss-120b $0.47Sonar Reasoning Pro $6.20
RAG over documents20K in, 800 outGPT-oss-120b $3.48Sonar Reasoning Pro $46.40
Coding agent step40K in (80% cached), 1.5K outGPT-oss-120b $4.50Sonar Reasoning Pro $92.00

Use both, with automatic fallback

Same key, same endpoint: change the model id between "gpt-oss-120b" and "sonar-reasoning-pro".

curl https://api.ultragpt.pro/v1/chat/completions \
  -H "Authorization: Bearer $ULTRAGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-oss-120b",
    "models": ["sonar-reasoning-pro"],
    "messages": [{ "role": "user", "content": "Hello!" }]
  }'

FAQ

Which is cheaper, GPT-oss-120b or Sonar Reasoning Pro?

GPT-oss-120b: $0.15 in / $0.6 out per 1M. Sonar Reasoning Pro: $2 in / $8 out per 1M. Prices are in USD and billed per use from your UltraGPT balance.

Can I use GPT-oss-120b and Sonar Reasoning Pro with the same API key?

Yes. Both are available through the OpenAI-compatible UltraGPT API at https://api.ultragpt.pro/v1: set model to "gpt-oss-120b" or "sonar-reasoning-pro".

Which has a bigger context window, GPT-oss-120b or Sonar Reasoning Pro?

GPT-oss-120b: 131,072 tokens. Sonar Reasoning Pro: 128,000 tokens.

More comparisons