# LLMoney

> LLM Token Cost Calculator & Model Pricing API.
> Compare pricing for 66 AI models from OpenAI, Anthropic, Google, DeepSeek, Meta, Mistral AI, xAI, Cohere, GigaChat (Cloud.ru), Yandex AI Studio.

## API Base URL

https://api.llmoney.ru/api/v1

All endpoints return JSON. No authentication required for read-only endpoints.

## Endpoints

### GET /api/v1/models

List all supported LLM models with current pricing.

Query parameters:
- provider (string, optional): Filter by provider slug (see Providers section)
- free_only (boolean, optional): Only show models with free (local) tokenization. Default: false

Example request:
  GET https://api.llmoney.ru/api/v1/models?provider=openai

Example response:
  {
    "models": [
      {
        "id": "openai-gpt-4o",
        "provider": "openai",
        "provider_display_name": "OpenAI",
        "name": "gpt-4o",
        "display_name": "GPT-4o",
        "tokenizer_type": "tiktoken",
        "input_price": 2.50,
        "output_price": 10.00,
        "cached_input_price": 1.25,
        "currency": "usd",
        "context_window": 128000,
        "description": null
      }
    ]
  }

Prices are per 1,000,000 tokens. Currency is "usd" or "rub".

### GET /api/v1/models/{model_id}

Get details for a specific model including current pricing.

Example request:
  GET https://api.llmoney.ru/api/v1/models/openai-gpt-4o

Example response:
  {
    "id": "openai-gpt-4o",
    "provider": "openai",
    "provider_display_name": "OpenAI",
    "name": "gpt-4o",
    "display_name": "GPT-4o",
    "tokenizer_type": "tiktoken",
    "tokenizer_config": {"encoding": "o200k_base"},
    "input_price": 2.50,
    "output_price": 10.00,
    "cached_input_price": 1.25,
    "currency": "usd",
    "is_paid_api": false,
    "supports_byok": true,
    "context_window": 128000,
    "description": null
  }

### GET /api/v1/models/{model_id}/price-history

Get historical pricing data for a model, sorted chronologically.

Example request:
  GET https://api.llmoney.ru/api/v1/models/openai-gpt-4o/price-history

Example response:
  {
    "model_id": "openai-gpt-4o",
    "entries": [
      {
        "date": "2025-01-01",
        "input_price": 2.50,
        "output_price": 10.00,
        "cached_input_price": 1.25,
        "source": "seed"
      }
    ]
  }

### GET /api/v1/recommend

AI-agent-friendly endpoint for selecting optimal models by budget and constraints.
Returns a simplified, sorted list without tokenizer details.

Query parameters:
- max_input_price (float, optional): Maximum price per 1M input tokens
- max_output_price (float, optional): Maximum price per 1M output tokens
- min_context_window (integer, optional): Minimum context window size in tokens
- currency (string, optional): Filter by currency — "usd" or "rub"
- provider (string, optional): Filter by provider slug
- sort_by (string, optional): Sort criterion — "input_price" (default), "output_price", "context_window"
- limit (integer, optional): Maximum number of results, 1-100. Default: 10

Example request:
  GET https://api.llmoney.ru/api/v1/recommend?max_input_price=5&currency=usd&sort_by=input_price&limit=5

Example response:
  {
    "models": [
      {
        "id": "openai-gpt-4o-mini",
        "provider": "openai",
        "name": "gpt-4o-mini",
        "display_name": "GPT-4o Mini",
        "input_price": 0.15,
        "output_price": 0.60,
        "cached_input_price": 0.075,
        "currency": "usd",
        "context_window": 128000
      }
    ],
    "total_available": 42,
    "filters_applied": {
      "max_input_price": 5,
      "currency": "usd",
      "sort_by": "input_price",
      "limit": 5
    }
  }

### POST /api/v1/calculate

Calculate token counts and costs for a given text across models,
including prompt-cache discounts and tool-use token overhead.

Request body (JSON):
  {
    "text": "Your text to analyze",
    "model_ids": ["openai-gpt-4o"],
    "output_multiplier": 1.0,
    "cached_input_ratio": 0.0,
    "tools": [
      {"type": "function", "count": 3, "calls": 2}
    ]
  }

- text (string, required): The text to tokenize and price
- model_ids (array of strings, optional): Restrict calculation to these models. Default: all models with free local tokenization
- output_multiplier (float, optional): Multiplier for estimated output tokens. Default: 1.0
- cached_input_ratio (float, optional): Share of input tokens served from the prompt cache, 0..1. Models with a cached_input_price get a blended input cost. Default: 0.0
- tools (array, optional): Tool-use token overhead estimation. Each item:
  - type (string, required): "code_interpreter" | "web_search" | "file_search" | "mcp" | "function"
  - count (integer, optional): Number of tools of this type defined. Default: 1
  - calls (integer, optional): Invocations per request. Default: 1
  - def_tokens / call_tokens / result_tokens (integer, optional): Override the built-in per-definition / per-call / per-result token estimates

Example request:
  POST https://api.llmoney.ru/api/v1/calculate
  Content-Type: application/json
  {"text": "Hello, how much does this cost?", "output_multiplier": 1}

Example response:
  {
    "input_length": 31,
    "word_count": 6,
    "results": [
      {
        "model_id": "openai-gpt-4o",
        "provider": "openai",
        "model_name": "gpt-4o",
        "display_name": "GPT-4o",
        "tokens": 8,
        "cached_tokens": 0,
        "input_cost": 0.00002,
        "cached_input_cost": 0.0,
        "output_cost": 0.00008,
        "tool_input_tokens": 0,
        "tool_output_tokens": 0,
        "tool_cost": 0.0,
        "total_cost": 0.0001,
        "currency": "usd",
        "tokenizer_used": "o200k_base",
        "is_available": true
      }
    ]
  }

### POST /api/v1/tokenize

Tokenize text using a specific model's tokenizer. Returns individual token IDs.

Request body (JSON):
  {
    "text": "Text to tokenize",
    "model_id": "openai-gpt-4o"
  }

### GET /api/v1/exchange-rate

Current USD to RUB exchange rate (Central Bank of Russia, cached server-side).

## Pricing Format

- All prices are per 1,000,000 (one million) tokens
- currency field is "usd" (US dollars) or "rub" (Russian rubles)
- input_price: cost per 1M input/prompt tokens
- output_price: cost per 1M output/completion tokens
- cached_input_price: discounted rate for cached input tokens (if available)
- Prices refresh automatically every day from provider sources; history is available via /price-history

## Providers

Available provider slugs: anthropic, cohere, deepseek, gigachat, google, meta, mistral, openai, xai, yandex

## OpenAPI Specification

Full machine-readable OpenAPI 3.1 spec:
  https://api.llmoney.ru/api/v1/openapi.json

Interactive Swagger UI documentation:
  https://api.llmoney.ru/api/docs

ReDoc documentation:
  https://api.llmoney.ru/api/redoc