PricePerToken Model Pricing

Get pricing for all LLM models with optional filtering.

Args: author: Filter by model author/creator (e.g., "openai", "anthropic", "google") min_context: Minimum context length (e.g., 100000 for 100k+) max_ttft: Maximum time to first token in seconds (e.g., 2.0 for under 2 seconds) min_tokens_per_second: Minimum output speed (e.g., 50 for 50+ tok/s) supports_vision: Filter for vision-capable models supports_reasoning: Filter for reasoning/thinking models limit: Maximum results (default 100)

Get detailed pricing and info for a specific model.

Args: slug: Model slug (e.g., "openai-gpt-4o"). Use this OR author+model. author: Model author (e.g., "openai"). Use with 'model' param. model: Model name (e.g., "gpt-4o"). Use with 'author' param. provider: Optional API provider for provider-specific pricing (e.g., "bedrock", "openrouter", "groq", "together", "fireworks", "deepinfra"). If omitted, returns the model author's direct pricing. If set, returns that provider's pricing instead.

Side-by-side comparison of multiple models.

Args: slugs: Comma-separated list of model slugs (e.g., "openai-gpt-4o,anthropic-claude-3-5-sonnet,google-gemini-2-0-flash") provider: Optional API provider for provider-specific pricing (e.g., "bedrock", "openrouter", "groq"). If omitted, uses each model author's direct pricing. Useful for comparing what models cost on a specific platform.

Returns: Comparison table with pricing, benchmarks, and capabilities for each model.

Compare pricing for the same model across different API providers.

Shows what each provider (Bedrock, OpenRouter, Groq, etc.) charges for a model, so you can find the cheapest place to run it.

Args: slug: Model slug (e.g., "openai-gpt-4o"). Use this OR author+model. author: Model author (e.g., "openai"). Use with 'model' param. model: Model name (e.g., "gpt-4o"). Use with 'author' param.

Returns: Pricing from every provider that offers this model, sorted by input price.

Get models ranked by a specific benchmark.

Args: benchmark: Benchmark to sort by. Options:

• "coding": Composite coding score (code generation, understanding, problem-solving)
• "intelligence": Composite score for general reasoning and knowledge
• "math": Composite mathematical reasoning score
• "aider": Aider Polyglot - real-world code editing across multiple languages
• "livecodebench": LiveCodeBench - competitive programming (Codeforces, LeetCode, AtCoder)

List all model providers/authors with model counts and price ranges.

Returns: List of providers with their model counts and min/max pricing.

Search for models by name, author, or description.

Args: query: Search term (e.g., "gpt", "claude", "vision", "coding") limit: Maximum results (default 20)

Returns: Matching models with pricing info.

Look up the provider-specific model ID/slug for any API provider.

Use this when you need the exact model identifier to pass to a provider's API (e.g., OpenRouter, Bedrock, Groq, Fireworks, Cerebras, Together, DeepInfra, SambaNova, Gemini).

Args: query: Natural language model name (e.g., "claude opus 4.6", "gpt 5", "gemini 2.5 pro"). If no results, try a shorter name (e.g., "claude opus" instead of "claude opus 4.6"). provider: Optional provider to filter by (e.g., "groq", "bedrock", "openrouter").