Model Differences

Positioning of the Three Model Lines

Anthropic offers Claude across three model lines: Opus, Sonnet, and Haiku. The names are derived from poetic forms and symbolize the weight of each model's capabilities.

Opus — Highest Performance, Highest Cost

This model is suited for the most complex reasoning, long-form comprehension, and advanced analytical tasks. The latest version is Claude Opus 4.8 (API ID: claude-opus-4-8). API pricing is $5 per 1M input tokens and $25 per 1M output tokens (USD, excluding tax; reference values), a significant reduction from previous generations. The context window is an exceptionally wide 1 million tokens, making it ideal for quality-first tasks such as legal document review and research summarization rather than high-volume batch processing.

Sonnet — Balanced, Flagship Model

This model strikes an excellent balance between speed and performance, and is the standard model available on claude.ai's free and Pro plans. The latest version is Claude Sonnet 4.6 (API ID: claude-sonnet-4-6). API pricing is $3 per 1M input tokens and $15 per 1M output tokens. The context window is 1 million tokens. It delivers strong cost-performance for coding, writing, and everyday question-answering. For most users, Sonnet will be more than sufficient in the vast majority of cases.

Haiku — Fastest, Most Affordable

This model offers the fastest response times and the lowest API pricing. The latest version is Claude Haiku 4.5 (API ID: claude-haiku-4-5). API pricing is $1 per 1M input tokens and $5 per 1M output tokens. The context window is 200k tokens. It is well-suited for throughput-intensive use cases such as chatbots, automated classification, and large-scale document processing.

How to Read Version Numbers

Model names follow the format Claude [Line] [Version]. For example, "Claude Sonnet 4.6" means "version 4.6 of the Sonnet line." Higher numbers indicate newer, more capable versions. Since Anthropic releases new versions regularly, we recommend checking the model comparison article for the latest information throughout and beyond this lesson.

Which Model Should You Choose?

The key decision framework is prioritizing among "quality vs. cost vs. speed."

• Quality first (legal, medical, high-precision analysis): Opus
• Balance-focused (development, writing, everyday tasks): Sonnet
• Cost and speed first (high-volume processing, bots, real-time responses): Haiku

In the claude.ai interface, you can select a model in the settings. When using the API, specify the model name directly in the model parameter.