Skip to main content

​
Available Models

Rose offers two model tiers for agents:

​
rose-1-lite

Best for: Fast tasks, simple research, web searches, basic automation
MetricValue
Input tokens$0.50 / 1M tokens
Output tokens$1.50 / 1M tokens
Use cases:
  • Quick web searches
  • Simple data lookups
  • Basic summarization
  • Fast task completion

​
rose-1-ultra

Best for: Complex reasoning, detailed analysis, multi-step research
MetricValue
Input tokens$3.00 / 1M tokens
Output tokens$15.00 / 1M tokens
Use cases:
  • Deep research and analysis
  • Complex multi-step tasks
  • Code generation and review
  • Detailed report writing

​
Specifying a Model

Pass the model parameter when creating an agent: 
curl -X POST "https://navi-j9a9.onrender.com/api/agents" \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{
    "goal": "Research quantum computing startups",
    "model": "rose-1-lite"
  }'
If no model is specified, rose-1-lite is used by default.

​
Choosing the Right Model

If you need…Use
Fast, cheap resultsrose-1-lite
High quality outputrose-1-ultra
Simple lookupsrose-1-lite
Complex analysisrose-1-ultra

​
Token Usage

Token usage is tracked per agent and visible in the status response: 
{
  "agent": {
    "agent_id": "...",
    "token_count": 15420,
    "status": "COMPLETED"
  }
}
Monitor your usage in the Dashboard to manage costs.