Claude 4 Sonnet

Trending

Anthropic · Active · Updated May 15, 2026

Anthropic's most balanced frontier model, delivering near-opus intelligence at significantly lower cost. Excels at coding, reasoning, and nuanced instruction following.

PricingCompare
Input Price
$3.00/M
per million tokens
Output Price
$15.00/M
per million tokens
Context Window
200,000
tokens
Max Output
8,192
tokens

Technical Specifications

ProviderAnthropic
Release DateMarch 15, 2025
Pricing Typeper token
Input Price$3.00 / 1M tokens
Output Price$15.00 / 1M tokens
Cached Input$0.30 / 1M tokens
Context Window200,000 tokens
Max Output8,192 tokens
Input Modalitiestext, image
Output Modalitiestext
Statusactive
Availabilityapi, web_app
Latencyfast
Rate Limit5,000 RPM
Pricing URLView official pricing →
Docs URLView documentation →

Capability Scores

Coding
94
Reasoning
90
Math
88
Image
72
Speed
78

Overview

Claude 4 Sonnet represents Anthropic's latest advancement in balancing intelligence with efficiency. It achieves near-identical performance to Claude 4 Opus on coding benchmarks while costing significantly less. The model excels at complex multi-step reasoning, code generation, and maintaining coherence across very long documents thanks to its 200K context window. For developers building AI-powered applications, Claude 4 Sonnet offers the best price-to-performance ratio in Anthropic's lineup.

Pros

  • +Top-tier coding performance (94/100) — leads the industry on HumanEval and SWE-bench
  • +200K context window handles large codebases and documents
  • +Excellent instruction following and safety alignment
  • +Cached input pricing at $1.50/M tokens reduces costs for repeated contexts

Cons

  • −No audio input modality — cannot process speech directly
  • −Output capped at 8K tokens, limiting very long generations
  • −Slower than GPT-4o mini and other lightweight models for simple tasks

Compare with Alternatives

vs Gpt 5 4

Claude 4 Sonnet wins for coding and reasoning; GPT-5.4 wins for speed, multimodal versatility, and value.

vs Deepseek v4 pro

DeepSeek V4 Pro wins for value, context window, and math; Claude 4 Sonnet wins for coding and ecosystem maturity.

Use Cases

Complex code generation and refactoring across large codebases
Advanced document analysis and summarization
Multi-step reasoning workflows and agentic applications

Frequently Asked Questions about Claude 4 Sonnet

How much does Claude 4 Sonnet cost?
Claude 4 Sonnet costs $3 per million input tokens and $15 per million output tokens. Cached input is $0.3 per million tokens.
What is the context window of Claude 4 Sonnet?
Claude 4 Sonnet has a 200,000 token context window, with a maximum output of 8,192 tokens.
Is Claude 4 Sonnet good for coding?
Claude 4 Sonnet scores 94/100 on coding benchmarks.
What modalities does Claude 4 Sonnet support?
Claude 4 Sonnet supports text, image input and text output.
← Back to all models