Claude Model Comparison: Haiku vs Sonnet vs Opus

Compare Claude Haiku 4.5, Sonnet 4.6, and Opus 4.6 side by side. Pricing, context windows, output speed, and use cases — everything you need to pick the right model for your project.

Filter by need:

Claude Haiku 4.5

Fastest & cheapest

Intelligence

Pricing / 1M tokens

$0.80

Input

$4.00

Output

200K

Context window

~80 tok/s

Output speed

Best for

High-volume API workloads
Real-time chat interfaces
Classification & routing
Simple Q&A and lookup
Background processing

Features

Vision / image input

Tool use / function calling

Streaming

Batch API

Extended thinking

Claude Code default

Claude Sonnet 4.6

Best balance

Intelligence

Pricing / 1M tokens

$3.00

Input

$15.00

Output

200K

Context window

~45 tok/s

Output speed

Best for

Production AI features
Code generation & review
Claude Code (default model)
Document analysis
Agentic workflows

Features

Vision / image input

Tool use / function calling

Streaming

Batch API

Extended thinking

Claude Opus 4.6

Most capable

Intelligence

Pricing / 1M tokens

$15.00

Input

$75.00

Output

200K

Context window

~25 tok/s

Output speed

Best for

Complex research synthesis
Advanced reasoning & math
Long-document understanding
Difficult code challenges
High-stakes outputs

Features

Vision / image input

Tool use / function calling

Streaming

Batch API

Extended thinking

Pricing as of early 2025. Always verify at anthropic.com/pricing.

Quick comparison

Model	Input / 1M	Output / 1M	Context	Speed
Claude Haiku 4.5Fastest & cheapest	$0.80	$4.00	200K	~80 tok/s
Claude Sonnet 4.6Best balance	$3.00	$15.00	200K	~45 tok/s
Claude Opus 4.6Most capable	$15.00	$75.00	200K	~25 tok/s

Browse MCPs for Claude

Extend any Claude model with tools — file access, web search, databases, and more.

Browse MCPs

How to choose a Claude model

Use Haiku 4.5 for high-volume, latency-sensitive, or budget-constrained workloads — classification, routing, simple Q&A
Use Sonnet 4.6 for most production use cases — it is the default model in Claude Code and the best balance of capability and cost
Use Opus 4.6 for the most complex tasks — long-form research, advanced reasoning, and high-stakes outputs where quality outweighs cost

About the Claude API

All three models are available through the Anthropic API and support vision (image input), tool use, and streaming. Opus 4.6 additionally supports extended thinking for step-by-step reasoning. Pricing is per million tokens for both input and output.