Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
199 lines (151 loc) · 5.98 KB

CONTRIBUTING.md

File metadata and controls

199 lines (151 loc) · 5.98 KB

Contributing to Portkey Models

Thank you for your interest in contributing to the Portkey Models repository! This community-maintained database helps developers track AI model pricing and configurations across 35+ providers.

Table of Contents

Getting Started

  1. Fork this repository
  2. Clone your fork locally
  3. Make your changes
  4. Test your changes (validate JSON)
  5. Submit a pull request

Types of Contributions

Contribution Directory Description
Model Pricing pricing/ Token costs, image pricing, additional units
Model Configs general/ Parameters, capabilities, supported features

Adding or Updating Pricing

Pricing files are located in pricing/{provider}.json.

Price Format

⚠️ Prices are in cents per token, not dollars.

Provider Pricing JSON Value Calculation
$0.03 / 1K tokens 0.003 $0.03 ÷ 1000 = 0.00003 dollars = 0.003 cents
$2.50 / 1M tokens 0.00025 $2.50 ÷ 1,000,000 = 0.0000025 dollars = 0.00025 cents
$10 / 1M tokens 0.001 $10 ÷ 1,000,000 = 0.00001 dollars = 0.001 cents

Quick formula: price_in_cents = (dollars_per_million / 1,000,000) * 100

Schema Structure

{
  "model-name": {
    "pricing_config": {
      "pay_as_you_go": {
        "request_token": { "price": 0.001 },
        "response_token": { "price": 0.002 },
        "cache_write_input_token": { "price": 0 },
        "cache_read_input_token": { "price": 0.0005 }
      },
      "currency": "USD"
    }
  }
}

Available Fields

Field Description
request_token Input token cost
response_token Output token cost
cache_write_input_token Cache write cost
cache_read_input_token Cache read cost
request_audio_token Audio input cost
response_audio_token Audio output cost
additional_units Provider-specific (web_search, thinking_token, etc.)
image Image generation pricing by quality/size

Adding a New Provider

  1. Create pricing/{provider-name}.json
  2. Use lowercase with hyphens (e.g., together-ai.json)
  3. Include a default entry with zero prices
  4. Add model-specific pricing
{
  "default": {
    "pricing_config": {
      "pay_as_you_go": {
        "request_token": { "price": 0 },
        "response_token": { "price": 0 }
      },
      "currency": "USD"
    }
  },
  "model-name": {
    "pricing_config": {
      "pay_as_you_go": {
        "request_token": { "price": 0.001 },
        "response_token": { "price": 0.002 }
      }
    }
  }
}

Adding Model Configurations

Configuration files are located in general/{provider}.json.

Schema Structure

{
  "model-name": {
    "params": [
      {
        "key": "max_tokens",
        "maxValue": 16384
      },
      {
        "key": "temperature",
        "defaultValue": 1,
        "minValue": 0,
        "maxValue": 2
      }
    ],
    "type": {
      "primary": "chat",
      "supported": ["tools", "image"]
    }
  }
}

Type Values

Primary Type Description
chat Chat completion models
text Text completion models
embedding Embedding models
image Image generation models
audio Audio models
Supported Features Description
tools Function calling support
image Vision/image input support
cache_control Prompt caching support

Submitting Your Changes

Before Submitting

  1. Validate JSON syntax

    # Validate all pricing files
for file in pricing/*.json; do jq empty "$file" && echo "✅ $file"; done

# Validate all general files
for file in general/*.json; do jq empty "$file" && echo "✅ $file"; done

  2. Check your pricing math — Remember: cents per token, not dollars!

  3. Find the source — You'll need to provide a link to official pricing

Pull Request Guidelines

  1. Title format: [provider] Brief description

    • Example: [openai] Add GPT-5 pricing
    • Example: [anthropic] Update Claude 4 cache pricing

  2. Include source link — Link to the provider's official pricing page

  3. One provider per PR — Makes review easier

  4. Sign the CLA — First-time contributors must sign our Contributor License Agreement

What Happens Next

  1. Automated checks validate your JSON
  2. A maintainer reviews your PR
  3. Once merged, changes sync to Portkey's systems

Code of Conduct

Please read our Code of Conduct before contributing. We're committed to providing a welcoming and inclusive environment.

Questions?

  • 💬 Discord — Ask in #open-source
  • 🐛 Issues — Report bugs or request features
  • 📖 Documentation — Learn more about Portkey

Thank you for helping keep AI model pricing accurate and up-to-date! 🙏