Model
Claude
Anthropic's family of AI models, first released in March 2023, now spanning the Fable, Mythos, Opus, Sonnet and Haiku lines.
Claude is the family of large language models developed by Anthropic. The first Claude was released on 14 March 2023 as a “next-generation AI assistant” after a closed alpha with partners including Notion, Quora and DuckDuckGo, and was offered through a chat interface and a developer API. A lighter, faster companion model, Claude Instant, launched alongside it.
The family has since been organized into tiers. As of October 2026 Anthropic lists five lines: Fable, Mythos, Opus, Sonnet and Haiku. Fable is the generally available frontier line; Anthropic describes the current Fable release as “a Mythos-level model” with added safeguards for cybersecurity and biology. Mythos is the same underlying class of model offered without those safeguards to a small set of vetted cyberdefence and life-sciences organizations through verification programs. Opus, Sonnet and Haiku span the range from highest capability to fastest and cheapest, and all three received new releases in September and October 2026.
Claude models accept text and images, including documents, charts, spreadsheets and code, and produce text, code and files such as documents and slides. Anthropic’s consumer apps add voice conversation and web search. The models are available in the Claude web, desktop and mobile apps; through Claude Code for developers; through the Claude Platform API; and on Amazon Bedrock, Google Cloud and Microsoft Foundry.
Access to Fable and Mythos was briefly suspended for some users in June 2026 amid US national security review and restored over the following weeks. Anthropic does not describe any Claude model as a Super Intelligence.
For current models, capabilities and pricing see anthropic.com.