Claude

Anthropic's flagship LLM family.

What is Claude?

Anthropic’s flagship LLM family. Best-in-class for tool use, long-context reasoning, and writing quality. We use Claude as the default model in the bootcamp.

Key Features

Best-in-class tool use accuracy + JSON-mode output
200K context window with strong long-context recall
Vision input — analyse images, charts, screenshots
Prompt caching cuts cost up to 90% on repeated prefixes
Native Computer Use for browser/desktop automation
Available via Anthropic API, AWS Bedrock, Google Vertex

FAQ

Claude vs GPT-5 in production? +

For tool use, long-context document Q&A, and writing quality, Claude Sonnet is our default. For image generation or DALL-E features, GPT wins. For raw cost-per-token, both are similar at the small-model tiers.

Which Claude tier should I use? +

Sonnet is the price/performance sweet spot for production agents — what we default to in the bootcamp. Opus for very long-context reasoning, Haiku for high-volume cheap calls.