Happy Horse 1.0 is now on ModelsLab

Try Now
Skip to main content
Available now on ModelsLab · Language Model

Anthropic: Claude Sonnet 4.5Anthropic Claude Sonnet 4.5

Try Anthropic: Claude Sonnet 4.5API Documentation
Anthropic: Claude Sonnet 4.5

Code Agents Autonomously

Top Coding

77.2% SWE-Bench Score

Handles 30+ hour autonomous coding sessions on complex projects.

Hybrid Reasoning

200K Token Context

Toggles fast execution and extended thinking for logic challenges.

Computer Control

61.4% OSWorld

Analyzes screenshots, executes clicks and navigation directly.

Examples

See what Anthropic: Claude Sonnet 4.5 can create

Copy any prompt below and try it yourself in the playground.

Code Database App

“Build a full-stack web app with PostgreSQL database, user authentication, REST API endpoints, and deployment script using Node.js and React. Include error handling, security best practices, and Docker configuration for production.”

Agent Workflow

“Design an autonomous agent that monitors system logs, detects anomalies using statistical analysis, generates alerts via email/Slack, and auto-remediates common issues like high CPU by scaling resources. Use Python with tool calls.”

Math Proof

“Prove the infinitude of primes using contradiction, then extend to Dirichlet's theorem on arithmetic progressions. Provide step-by-step reasoning with formal notation and verify with numerical examples up to 1000.”

Refactor Codebase

“Analyze this 10K-line Python codebase for performance bottlenecks, refactor using async/await patterns, add type hints, unit tests with pytest, and optimize database queries. Output diff and explanation.”

For Developers

A few lines of code.
Agents. Three lines.

ModelsLab handles the infrastructure: fast inference, auto-scaling, and a developer-friendly API. No GPU management needed.

  • Serverless: scales to zero, scales to millions
  • Pay per token, no minimums
  • Python and JavaScript SDKs, plus REST API
API Documentation 
import requests


response = requests.post(
    "https://modelslab.com/api/v7/llm/chat/completions",
    json={
  "key": "YOUR_API_KEY",
  "prompt": "",
  "model_id": ""
}
)
print(response.json())

FAQ

Common questions about Anthropic: Claude Sonnet 4.5

Read the docs

Anthropic: Claude Sonnet 4.5 model is a balanced LLM excelling in coding and agents. It scores 77.2% on SWE-bench Verified. Supports 200K context and hybrid reasoning.

Costs $3 per million input tokens and $15 per million output tokens. Matches prior Sonnet pricing for cost-effective heavy use. Tool use optimizes token consumption.

Leads in autonomous 30+ hour coding and computer use at 61.4% OSWorld. Includes memory across sessions and context editing. Most aligned model with reduced risks.

Yes, builds complete apps including databases, DNS setup, and security audits. Handles multi-file edits and long-horizon planning. Integrates with VS Code and Agent SDK.

Tool handling, cross-conversation memory, and smart context management. Clears old tool results automatically. Excels in finance, law, medicine, and STEM reasoning.

Improved safety profile reduces sycophancy and prompt injections. Defends agentic risks effectively. Outperforms prior models in alignment benchmarks.

Ready to create?

Start generating with Anthropic: Claude Sonnet 4.5 on ModelsLab.

Try Anthropic: Claude Sonnet 4.5API Documentation