# Anthropic API — AI Compare Ninja

> Developer API for Claude models with token-based billing, separate from Claude subscriptions.

## Overview
- **Category:** Development & Code (Agent & RAG frameworks)
- **Status:** Active · checked - Active.
- **Paid from:** $2 / $10 / per 1M input/output tokens (Claude Sonnet 5.5)
- **Free access:** No free plan published
- **Website:** https://platform.claude.com/
- **Canonical URL:** https://www.aicompare.ninja/en/tool/anthropic-api

## What it does
- Sonnet 5.5: $2 input and $10 output per million tokens; Sonnet 5 stays the same
- Opus 5.5: $4 / $20; Opus 5: $5 / $25; Haiku 4.5: $1 / $5 per million tokens
- Mythos 5.1 (invite only): $10/$50 per million tokens, cache read $0.25
- Managed Agents runtime: $0.08 per session-hour plus model tokens

## In brief
The Anthropic API provides Claude models for custom applications and bills per token. This entry keeps API rates clearly separate from the already-listed Claude consumer subscriptions.

Claude Managed Agents (hosted agent sessions) bills the same model token rates plus $0.08 per session-hour, charged only while the session status is "running".

Claude Mythos 5.1 is invite-only (Project Glasswing) and costs the same as Fable 5.1: $10 input and $50 output per million tokens; cache reads $0.25 per million tokens, batch $5/$25.

Claude Opus 5.5 is on the token price list at $4 input and $20 output per million tokens; Opus 5 stays $5 / $25.

Claude Sonnet 5.5 is listed under Latest models at $2 input and $10 output per million tokens. Sonnet 5 stays under Legacy at the same price.


## Limits
API billing is separate from Claude consumer subscriptions. For Fable 5.1, cache reads cost $0.25 per million tokens.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Claude Sonnet 5.5 | $2 / $10 | per 1M input/output tokens | $2 input and $10 output per MTok. Listed under Latest models.; Cache writes: $2.50 (5 min) / $4 (1 hr) per MTok; cache reads: $0.10 per MTok. Sonnet 5 stays under Legacy at $2 / $10. |
| Claude Sonnet 5 | $2 / $10 | per 1M input/output tokens | $2 input and $10 output per MTok; Standard price; the announced increase will not occur |
| Claude Fable 5.1 | $10 / $50 | per 1M input/output tokens | $10 input and $50 output per million tokens; Cache reads: $0.25 per million tokens |
| Claude Opus 5.5 | $4 / $20 | per 1M input/output tokens | $4 input and $20 output per MTok. Opus 5 stays $5 / $25 and is not replaced.; Cache writes: $5 (5 min) / $8 (1 hr) per MTok; cache reads: $0.20 per MTok. Fast mode (research preview, Claude API only): $8 / $40 across the full context window. |
| Claude Opus 5 | $5 / $25 | per 1M input/output tokens | $5 input and $25 output per MTok; Cache writes: $6.25 (5 min) / $10 (1 hr) per MTok; cache reads: $0.50 per MTok; Fast mode (research preview, Claude API only): $10 input / $50 output per MTok across the full context window |
| Claude Haiku 5.5 | $0.10 / $0.50 | per 1M input/output tokens | Prompts up to 100,000 tokens: $0.10 input and $0.50 output per MTok. Cache writes: $0.125 (5 min) / $0.20 (1 hr) per MTok; cache reads: $0.01 per MTok.; Prompts over 100,000 tokens: $0.50 input and $2.50 output per MTok. Cache writes: $0.625 (5 min) / $1 (1 hr) per MTok; cache reads: $0.05 per MTok.; Batch: $0.05 input and $0.25 output per MTok up to 100,000 tokens, and $0.25 input and $1.25 output per MTok above that. Haiku 4.5 stays at $1 / $5. |
| Claude Haiku 4.5 | $1 / $5 | per 1M input/output tokens | $1 input and $5 output per MTok; Cache writes: $1.25 (5 min) / $2 (1 hr) per MTok; cache reads: $0.10 per MTok |
| Claude Mythos 5.1 (invite only) | $10 / $50 | per 1M input/output tokens | $10 input and $50 output per million tokens; invite only / Project Glasswing; Cache reads: $0.25 per million tokens; batch $5/$25; limited availability, not GA |
| Claude Managed Agents Runtime | $0.08 | per session-hour | $0.08 per session-hour, only while status is running, metered in milliseconds; Model tokens at standard rates; container hours not billed on top; idle/rescheduling/terminated not charged |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://platform.claude.com/docs/en/about-claude/pricing
- https://platform.claude.com/docs/en/models/mythos-5-1/overview
- https://www.anthropic.com/claude-fable-and-mythos-5-1
- https://claude.com/pricing
- https://www.anthropic.com/pricing

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/anthropic-api*