Accurate pricing for 2,000+ models across 40+ providers. Free API, no auth required.

Explorer · Rankings · API · Contributing · Discord

This repository contains comprehensive pricing and configuration data for LLMs. It powers cost attribution for 200+ enterprises running 400B+ tokens through Portkey AI Gateway every day.

LLM pricing is a mess. Every team building on LLMs ends up maintaining its own spreadsheet — and it's stale by Friday.

• The Naming Problem — gpt-5, gpt-5.2-pro-2025-12-11, o1, o3-mini are all different models with different prices
• The Units Problem — OpenAI charges tokens, Google charges characters, Cohere uses "generations" and "summarization units"
• The Hidden Dimensions Problem — Thinking tokens, cache writes vs. reads, context thresholds, per-request fees, multimodal surcharges
• The Velocity Problem — DeepSeek dropped R1 pricing 75% in weeks. Google releases new tiers mid-quarter with no announcement

We built this as an open, community-maintained database to solve that problem.

🔓 Free API. No authentication required.

curl https://configs.portkey.ai/pricing/{provider}.json

# Examples
curl https://configs.portkey.ai/pricing/bedrock.json
curl https://configs.portkey.ai/pricing/openai.json
curl https://configs.portkey.ai/pricing/anthropic.json

curl https://api.portkey.ai/model-configs/pricing/{provider}/{model}

# Examples
curl https://api.portkey.ai/model-configs/pricing/openai/gpt-5
curl https://api.portkey.ai/model-configs/pricing/anthropic/claude-sonnet-4-5-20250514
curl https://api.portkey.ai/model-configs/pricing/google/gemini-3.0-pro

📖 Full API Documentation — Response schema, model configuration endpoints, and more.

⚠️ Prices are in cents per token, not dollars.

const costDollars = (tokens * price) / 100;

{
  "model-name": {
    "pricing_config": {
      "pay_as_you_go": {
        "request_token": { "price": 0.00025 },
        "response_token": { "price": 0.001 },
        "cache_write_input_token": { "price": 0 },
        "cache_read_input_token": { "price": 0.000125 },
        "additional_units": {
          "web_search": { "price": 1 }
        }
      },
      "currency": "USD"
    }
  }
}

// OpenAI (gpt-5, o3, o4-mini)
"additional_units": {
  "web_search": { "price": 1 },
  "file_search": { "price": 0.25 }
}

// Google (gemini-3.0-pro)
"additional_units": {
  "thinking_token": { "price": 0.001 },
  "web_search": { "price": 3.5 }
}

// OpenAI Sora
"additional_units": {
  "video_duration_seconds_720_1280": { "price": 10 }
}

// Together AI (image models)
"additional_units": {
  "megapixels": { "price": 0.05 }
}

Batch API pricing is defined in a separate batch_config section at the same level as pay_as_you_go. Prices are specified as exact values matching the provider's published batch pricing.

Schema:

{
  "pricing_config": {
    "pay_as_you_go": {
      "request_token": { "price": 0.00025 },
      "response_token": { "price": 0.001 }
    },
    "batch_config": {
      "request_token": { "price": 0.000125 },
      "response_token": { "price": 0.0005 }
    }
  }
}

Notes:

• Batch prices are typically 50% of standard pricing for text models
• Embedding models typically have 20% discount for batch

Supported Providers: OpenAI, Anthropic, Google (Vertex AI)

The easiest way to contribute is to pick an issue with the good first issue tag 💪.

1. Fork this repo
2. Edit pricing/{provider}.json
3. Submit a PR with source link

Remember: Prices are in cents per token: $0.03/1K → 0.003

Join our growing community around the world, for help, ideas, and discussions on AI.

• Chat with us on Discord
• Follow us on Twitter
• Connect with us on LinkedIn
• Visit us on YouTube
• Read our Blog
• View our official Documentation

Built by Portkey