Price Per Token
Compare LLM API pricing across 300+ models from OpenAI, Anthropic, Google, and 30+ providers. [Free]
Last verified:
What is Price Per Token?
Price Per Token is a free web tool that aggregates and compares LLM API pricing across 556+ AI models from 30+ providers including OpenAI, Anthropic, Google, Mistral, DeepSeek, and more. The tool displays up-to-date token costs (input and output prices per million tokens), context lengths, benchmark scores (Coding, MMLU, GPQA), and allows users to sort models by lowest cost to find the most affordable options for their projects.
Key features include a comprehensive pricing table sorted by input token cost, a cheapest LLM API page with 300+ models sorted by lowest cost, interactive pricing history charts showing historical trends, benchmark data from Artificial Analysis and HuggingFace, filtering by capability (coding, math, vision), and an MCP server for real-time pricing data directly in Claude Code, Cursor, and Windsurf. The site also offers a weekly newsletter on pricing changes and new releases.
This tool is ideal for AI developers, researchers, startup founders, and anyone building applications with LLM APIs who needs to compare costs and find the best value. It's especially useful for developers using AI coding assistants like Cursor, Cline, GitHub Copilot, and Aider who want to optimize their API spending while selecting models with appropriate performance characteristics.
Price Per Token pricing
Pricing model: Freemium
FREE - The tool is completely free to use with no paid plans. All features including the full pricing table, cheapest LLM API page, pricing history charts, benchmark data, and MCP server access are available at no cost. The MCP server specifically requires no API key and is free to use without authentication.
Price Per Token pros
- Completely free to use with no paid plans required
- Tracks 556+ LLM models from 30+ providers in one place
- 35 free models available including Gemma 3 1B and GLM-5 FP4
- Pricing data updated daily from official provider APIs and OpenRouter
- Shows both input and output token pricing for each model
- Includes benchmark scores (Coding, MMLU, GPQA) for performance comparison
- Context length displayed for every model
- Interactive pricing history charts show cost trends over time
- Cheapest LLM API page sorted by lowest input token cost
- MCP server integration for Claude Code, Cursor, and Windsurf
- No API key required for MCP server access
- Weekly newsletter on pricing changes and new model releases
- Filter by capability to find models for coding, math, or vision tasks
- Shows lowest price across all providers for each model
- Direct 'Try' links to test models immediately
Price Per Token cons
- No API access for programmatic integration beyond MCP server
- Pricing displayed is for prompts under 200k tokens only
- Some models use tiered pricing not fully shown in table
- No built-in cost calculator for custom prompt sizes
- Limited to API pricing, no self-hosted model cost comparison
- No user accounts or saved model preferences
- Benchmark data sourced from third parties not independently verified
- No mobile app available, web-only interface
Frequently asked questions about Price Per Token
What is the price range for LLM APIs on this site?
LLM API pricing ranges from free to $150.00 per million input tokens across 554+ models tracked on the page. There are 35 models available at no cost including Gemma 3 1B (Pretrained), GLM-5 FP4, and Hcompany/Holo3-35B-A3B. The cheapest paid LLM APIs include LFM2 24B A2B Preview at $0.00/M input tokens, Qwen3.5 0.8B at $0.01/M input tokens, and LiquidAI/LFM2-8B-A1B at $0.01/M input tokens.
What is the difference between input and output tokens?
Input tokens are what you send to the model (your prompt), and output tokens are what the model generates (the response). Output tokens are typically more expensive because generation is more computationally intensive. For example, Azure OpenAI charges $75.00/M input vs $150.00/M output tokens.
How often is the pricing data updated?
The pricing data is updated daily from official provider APIs and OpenRouter. The site tracks price changes across 554+ models, and historical pricing trends are available on the pricing history page with interactive charts showing percentage changes over time.
Does the MCP server require an API key?
No, the Price Per Token MCP server is free to use and doesn't require authentication. It provides real-time LLM pricing, benchmark, and speed data directly in Claude Code, Cursor, Windsurf, and other MCP-enabled AI assistants without any API key.
What providers are covered on this site?
The site covers 30+ providers including OpenAI, Anthropic, Google, Mistral AI, DeepSeek, Groq, Fireworks AI, Cerebras, Together AI, AWS Bedrock, Azure OpenAI, Google AI Studio, Nebius AI, Cloudflare Workers AI, OpenRouter, Cohere, Nvidia, Alibaba, Baidu, Tencent, Moonshotai, Z.ai, xAI, Perplexity, and many more.
What benchmarks are available for each model?
The site displays benchmark scores from Artificial Analysis and HuggingFace Open LLM Leaderboard including Coding scores, MMLU (Massive Multitask Language Understanding), and GPQA (Graduate-Level Google-Proof Q&A). Additional rankings exist for specific use cases like best LLM for coding, math, writing, RAG, and local LLM.
Can I find the cheapest model for my specific use case?
Yes, the Cheapest LLM API page shows 300+ AI models sorted by lowest input token cost. You can filter by capability (coding, math, vision) and see benchmark scores alongside pricing to find the best value. The table shows the lowest price across all providers for each model.
Are there free LLM models available?
Yes, there are 35 free models available at $0.00 per million tokens including Gemma 3 1B (Pretrained), GLM-5 FP4, Hcompany/Holo3-35B-A3B, Gemma 4 E2B IT, Gemma 3 270M IT, Llama 3.3 70B Instruct FP8 LoRA, Llama 4 Scout 17B 16E Instruct FP8 LoRA, Mixtral 8x7B Instruct FP8 LoRA, and many others from Google, Meta, and Mistral.
How do I set up the MCP server for Claude Code?
To set up the MCP server for Claude Code, run this command in your terminal: claude mcp add pricepertoken --transport http --url https://api.pricepertoken.com/mcp/mcp. For Cursor, add the server URL to ~/.cursor/mcp.json. For Claude Desktop, add it to claude_desktop_config.json. For Windsurf, add it to ~/.codeium/windsurf/mcp_config.json.
What tools and resources are available beyond the pricing table?
The site offers multiple tools including a Pricing Calculator, Token Counter, LLM Playground, AI Coding Tracker, MCP Server, and Subscription Calculator. It also has directories for AI Coding Assistants, AI Agent Frameworks, LLM Observability Tools, Local LLM Runners, and MCP Servers, plus rankings for best LLMs for specific tools like Cursor, Cline, Aider, GitHub Copilot, and Bolt.