Models overview (Opus 4.7, Sonnet 4.6, Haiku 4.5, deprecation policy)

The Claude model family is a set of large language models from Anthropic, organized into three named tiers—Opus, Sonnet, and Haiku—each balancing capability, speed, and cost differently. Opus is the most powerful tier, designed for complex reasoning, long-horizon autonomous tasks, and demanding software engineering. Sonnet sits in the middle, offering a practical balance of intelligence and throughput suitable for most production workloads. Haiku is the fastest and most cost-efficient tier, built for real-time applications and high-volume processing where latency and cost matter most. All current Claude models accept text and image inputs and produce text output. They support multilingual use and vision capabilities. Each model is identified by a string you pass in the API's `model` parameter, and starting with the Claude 4.6 generation, Anthropic moved to dateless model IDs (e.g., `claude-sonnet-4-6`) that act as pinned snapshots rather than rolling pointers. Anthropic maintains a formal model lifecycle with four stages: Active (fully supported and recommended), Legacy (no longer updated, may be deprecated), Deprecated (still functional but not recommended, with a named replacement and scheduled retirement date), and Retired (API requests return an error). Anthropic commits to at least 60 days of notice before moving a model to Retired status.