Claude API Pricing Explained: Rates and Model Costs (2026)

Update (9 Nov 2026): Claude Fable 5.2 and Mythos 5.2 are out - prices, access and benchmarks are in my Claude Fable 5.2 guide.

👁 113 reads

Understanding claude api pricing comes down to tracking token rates across different model tiers, from lightweight utility models to flagship reasoning systems. Official Anthropic pricing ranges from $0.10 to $10.00 per million input tokens and $0.50 to $50.00 per million output tokens depending on whether you run Haiku, Sonnet, Opus, or Fable. If you want a flexible setup with unified billing, you can check my service, Jantaris, which provides full SDK compatibility alongside media generation tools.

Official Claude API Pricing Across All Models

Anthropic structures its catalog into distinct tiers designed for speed, deep analytical thought, and code orchestration. When budgeting your claude api price, note that rates are measured in units of one million tokens (1M tokens). Below are the current official catalog rates:

Claude Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100k tokens, but it exists solely at Anthropic. If you are comparing landscape rates across competitors, you can also read about GPT-5.6 Sol API prices double or our tests in Gemini 4 Argon Test: 23 Tasks vs Opus 5.5 and GPT-6 Astra.

Anthropic API Pricing for a Typical Production Job

To visualize the actual cost of using the claude api, consider a standard production workload: executing 1,000 requests where each request consists of 2,000 input tokens and 500 output tokens. In aggregate, this equals exactly 2,000,000 input tokens (2M) and 500,000 output tokens (0.5M).

Using the official rates, here is what this workload costs across the model lineup:

• Claude Opus 5.5: (2M * $4.00) + (0.5M * $20.00) = $8.00 + $10.00 = $18.00
• Claude Sonnet 5.5: (2M * $2.00) + (0.5M * $10.00) = $4.00 + $5.00 = $9.00
• Claude Fable 5.1: (2M * $10.00) + (0.5M * $50.00) = $20.00 + $25.00 = $45.00
• Claude Opus 4.8: (2M * $5.00) + (0.5M * $25.00) = $10.00 + $12.50 = $22.50
• Claude Haiku 4.5: (2M * $1.00) + (0.5M * $5.00) = $2.00 + $2.50 = $4.50
• Claude Haiku 5.5: (2M * $0.10) + (0.5M * $0.50) = $0.20 + $0.25 = $0.45

You can run your own workload scenarios with an interactive cost calculator to forecast large-scale automation runs.

Caching and Context Window Considerations

When running long context inputs through the official Anthropic API, caching makes a massive difference in billing. Cache reads for Opus 5.5 and Sonnet 5.5 cost $0.20 per 1M tokens at Anthropic. If you repeatedly query documentation, large codebases, or complex system prompts, structuring your payloads to hit cache points dramatically cuts your monthly balance burn.

How to Buy Directly from Anthropic

To buy tokens directly from the model creator:

1. Navigate to the official Anthropic Console at claude.com/pricing.
2. Create your developer account and configure organization billing.
3. Generate your API credentials in the settings dashboard.

If you intend to use developer tools like Claude Code, note that Claude Code is included in Anthropic Pro, Max, Team, and Enterprise plans, but not in the Free plan. Detailed specifications for code CLI setups can be found in our overview of Kling 4.0 MCP and CLI tools and the official documentation at code.claude.com/docs.

Comparing Claude API Pricing with Jantaris

In addition to writing technical analyses on this blog, I built a service called Jantaris. It gives you access to chat models alongside image, video, and audio generation under a single prepaid balance in USD without recurring subscriptions. When I checked the price lists on 18 October 2026, my service offered lower rates on four specific Claude models compared to official pricing:

You can see the full developer rates on the pricing catalog page. Billing operates on a prepaid USD balance top-up by card with a $10 minimum ($10, $25, $50, $100, $250). No tax is added at payment, the full amount goes to the balance, there is no minimum commitment, and the balance never expires. Prompts and outputs are never used to train models, and failed generations cost $0.

SDK Integration Code

You do not need to rewrite your software to point to my service. It supports native SDK endpoints by updating the base URL parameter.

Using the official Anthropic Python SDK:

from anthropic import Anthropic
client = Anthropic(base_url="https://jantaris.com", api_key=os.environ["JANTARIS_API_KEY"])
msg = client.messages.create(model="claude-sonnet-5.5", max_tokens=1024, messages=[{"role": "user", "content": "..."}])

Using the OpenAI Python SDK:

from openai import OpenAI
client = OpenAI(base_url="https://jantaris.com/v1", api_key=os.environ["JANTARIS_API_KEY"])
resp = client.chat.completions.create(model="claude-sonnet-5.5", messages=[{"role": "user", "content": "..."}])

Configuring Claude Code in your terminal (macOS/Linux):

export ANTHROPIC_BASE_URL=https://jantaris.com
export ANTHROPIC_AUTH_TOKEN=sk-jt-...
export ANTHROPIC_MODEL=claude-sonnet-5.5
claude

For more setup details, see the quickstart documentation.

Frequently Asked Questions

What is the cheapest Claude model available via API?

Claude Haiku 5.5 is the lowest-cost model at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100k tokens. It is available exclusively through Anthropic’s direct API platform.

How does Claude API pricing calculate input versus output tokens?

Tokens are measured in increments of one million. Input tokens cover your prompt and system instructions, while output tokens account for text generated by the model, with output tokens priced significantly higher across all models.

Can I use Claude Code with alternative API endpoints?

Yes, you can direct Claude Code to alternative endpoints by defining the ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN environment variables in your terminal. This routes your CLI interactions seamlessly without requiring code modifications.

How much does Claude Sonnet 5.5 cost for 1,000 standard requests?

For 1,000 requests of 2,000 input tokens and 500 output tokens (2M input and 0.5M output tokens total), official Claude Sonnet 5.5 pricing comes to exactly $9.00. On my service, that exact volume costs $8.505.

Do API balances expire if left unused?

On the direct Anthropic developer platform, credit policies depend on your billing setup. On my service, your prepaid balance never expires and failed requests cost $0.

Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →

Sources

• Anthropic Official Pricing Table
• Claude Code Documentation
• OpenAI API Pricing Documentation
• Google Gemini API Pricing Reference

Prices checked on 18 October 2026.