Model Submission

Your API key will be encrypted and used only for benchmark testing.

API Requirements

Technical specifications

Endpoint Format

Open-AI compatible endpoint (see OpenAI Docs)

response = self.client.chat.completions.create( model=self.model, messages=messages, max_tokens=kwargs.pop("max_tokens", DEFAULT_MAX_TOKENS), **kwargs )

Request Body

{
  "role": "You are a contestant on The Price is Right...",
  "content": "Ajax dishwashing liquid..."
}

Expected Response

{
  "bid": 1234.56, 
  "rationale": "Brief explanation of how you estimated the value."
}

Review Process

  • Initial Validation: We'll test your endpoint with sample requests to ensure compatibility.
  • Benchmark Testing: Your model will compete in 50 showcase rounds.
  • Results & Publication: We'll add your model to the public leaderboard if it meets quality standards.

Important Notes

  • Your model must respond with valid JSON in the specified format
  • We may run approximately 50-100 requests for the full benchmark
  • API keys are encrypted and never shared publicly
  • API keys are encrypted and never shared
  • Models that don't follow format requirements may be excluded