16x EvalAnthropic offers a family of AI products and services under the Claude brand. However, the similar names for different offerings can be confusing, especially when used in 3rd party tools and services like Cline and Repo Prompt.
Let's take a look at the distinctions between Claude, Claude API, and Claude Code to help you understand what's the difference between them and which one should you use to get the most out of Claude without breaking the bank.
The Claude Product
When most people refer to Claude, they are talking about the consumer-facing product made by Anthropic, the company behind the Claude models. This includes the web interface at claude.ai, as well as desktop and mobile applications. This product allows you to chat with Claude models for various tasks like writing, summarizing, and analyzing information.
Anthropic offers several subscription plans for the Claude product. As outlined on their official pricing page, there is a free tier for basic use, a Pro plan for individuals, and Max plans for those who need more usage. These subscriptions provide access to more powerful models and features directly through the Claude applications.
Here is a simple breakdown of the individual plans:
| Plan | Monthly Price | Key Feature |
|---|---|---|
| Free | $0 | Basic access on web, iOS, and Android |
| Pro | $20 ($17 if billed annually) | More usage and access to Claude Code |
| Max | $100 or $200 | 5x or 20x more usage and early feature access |
The Claude API
When most people refer to Claude API, they are talking about accessing Claude models via Anthropic API. It is a service for developers who want to integrate Anthropic's Claude models into their own applications. Instead of using the Claude chat interface, developers make programmatic calls to models like Claude Opus 4.1 and Sonnet 4.
While the more precise term is Claude via Anthropic API, we will stick to Claude API as it is easier to understand.
If you are using Claude models in a 3rd party tool like Cursor, Cline or Repo Prompt, you are likely using Claude API. There are exceptions (Claude Code), but more on that in the next section.
Claude API Pricing
Unlike the subscription-based Claude product, Claude API is typically priced on a pay-per-use basis. You are charged based on the number of tokens you send (input) and receive (output).
There are two ways you can pay for the Claude API usage:
- You pay the API cost directly to Anthropic via their Bring-Your-Own-Key (BYOK) by entering your own API key into the 3rd party tool.
- Or you can pay the 3rd party tool for the usage without directly interacting with Claude API. For example, Cline and Cursor both operate as a provider (proxy service), which means you pay them for the usage, and they pay Anthropic for the usage.
Claude Code
Claude Code is a separate, specialized product by Anthropic designed specifically for software development. It is an agentic coding assistant that you can use directly in your terminal or integrated into an IDE. It is designed to help with tasks like building new features from a description, fixing bugs, and navigating complex codebases.
Claude Code can directly edit files, run commands, and create commits, making it an active participant in the development workflow. It is designed to be scriptable and can be composed with other command-line tools, following a Unix-like philosophy.
Claude Code Pricing
There are two ways to gain access to Claude Code, with different pricing: One way is through a subscription to the Claude Pro or Max plans. Claude Pro subscription or a Max subscription includes access to use Claude Code directly in your terminal.
The other way is to pay per usage via Claude API, where you pay for the API cost directly to Anthropic (as discussed in the previous section).
This can be more expensive than the subscription-based pricing if you are a heavy user, as the subscription offers more API usage for the price. For example, you can get $150 worth of API usage for the price of $20 a month with a Claude Pro subscription.
Claude Code SDK (Confusion Point)
Now we come to the most confusing part. Beside using Claude Code directly via terminal UI (TUI) as a user, Claude Code also offers programmatic access via its SDKs or a CLI tool.
This means that tools like Cline and Repo Prompt can integrate with a user's local Claude Code installation and gain access to Claude Code features without paying separate API fees.
Here's a diagram to help you understand how this works:
This means that if you have set up Claude Code locally with Claude subscription, you can use it in Cline and Repo Prompt without paying separate API fees.
Cline has documentation on how you can set this up. As of writing, Repo Prompt appears to not require any additional setup to use Claude Code.
Summary
Here's a summary of the differences between the three offerings:
| Type | Usage | Pricing | |
|---|---|---|---|
| Claude | Consumer-facing chat interface | The consumer-facing product made by Anthropic. | Free tier with limited usage Claude Pro ($20/month) Claude Max 5x ($100/month) Claude Max 20x ($200/month) |
| Claude API | Developer API service | A service for developers to integrate Claude models into their own applications. | Pay-per-use based on input/output tokens Or through 3rd party providers |
| Claude Code | Specialized development tool | A specialized product for software development. Also available as a CLI tool or SDK, which allows programmatic use of Claude Code features inside 3rd party tools. | Included with Claude Pro/Max subscriptions Or pay-per-use via API |
That's the end of the article. Hope you found it helpful in understanding the differences between Claude, Claude API, and Claude Code.
If you are building an AI product, you will likely need to do evaluations on the prompts and models. 16x Eval is a desktop app designed to help you run evals systematically, with a user-friendly GUI and a rich set of features.
You can use it to test how different models perform with the same prompt, or test different variations of the same prompt to see which one works best for your use case.