How Claude works.
Claude is Anthropic's model family for language, reasoning, coding, vision, and tool use. Anthropic currently lists Claude Fable 5, Claude Opus 5, Claude Sonnet 5, and Claude Haiku 4.5. The first three support context windows up to one million tokens, while Haiku is positioned for lower-latency work with a smaller context window.
Large context can simplify some document and code workflows, but it does not remove the need to select relevant evidence, evaluate outputs, and control cost. Choose among the tiers using a representative test set and the latency, context, and budget constraints of the actual job.
The lineup, platform availability, limits, and pricing can change. Check Anthropic's current model overview before committing to an implementation.
Where it helps.
- 01Long document analysis (1M context)
- 02Code review and generation
- 03Complex reasoning and analysis
- 04AI-assisted development (Claude Code)
- 05Enterprise knowledge work