This is Anthropic's announcement of Claude, published March 14, 2023. It is the primary record of the first public release of the Claude model family, which runs through to the 4.x and Mythos generations, and it is the document that first states publicly what the company claimed for the product at launch. Its classification as a foundational primary text, despite its announcement form, follows the same reasoning as ChatGPT: Optimizing Language Models for Dialogue (OpenAI, November 30, 2022).
The post opens by describing the preceding closed alpha: "After working for the past few months with key partners like Notion, Quora, and DuckDuckGo in a closed alpha, we've been able to carefully test out our systems in the wild, and are ready to offer Claude more broadly so it can power crucial, cutting-edge use cases at scale."
What the post claims for the model
Claude is described as "a next-generation AI assistant based on Anthropic's research into training helpful, honest, and harmless AI systems" — the formulation that connects the product to the company's published research programme; see Constitutional AI: Harmlessness from AI Feedback and Constitutional AI. It is stated to be accessible "through chat interface and API in our developer console" and "capable of a wide variety of conversational and text processing tasks while maintaining a high degree of reliability and predictability."
The named use cases are "summarization, search, creative and collaborative writing, Q&A, coding, and more." The claims attributed to early customers are comparative rather than measured: that Claude "is much less likely to produce harmful outputs, easier to converse with, and more steerable — so you can get your desired output with less effort," and that it "can also take direction on personality, tone, and behavior." No benchmark results, model sizes, training-compute figures or evaluation data appear in the post.
Two versions are announced: "Claude is a state-of-the-art high-performance model, while Claude Instant is a lighter, less expensive, and much faster option." The post commits to continued work "to make them more helpful, honest, and harmless as we learn more from our safety research and our deployments."
Launch partners
The bulk of the post is partner testimonials, which together record the first commercial deployments of the model:
- Quora, offering Claude through its Poe AI chat app. Autumn Besselman, Head of People and Comms, is quoted saying users "describe Claude's answers as detailed and easily understood, and they like that exchanges feel like natural conversation." Two unnamed Poe users are quoted, one saying "Claude feels more conversational than ChatGPT."
- Juni Learning, using Claude for its Discord Juni Tutor Bot. CEO Vivian Shen states the company "evaluated Anthropic against competitors" and chose Claude for its "helpful, high-quality responses," citing math problems and symbolism in critical reading as subjects where it "provided better, richer answers."
- Notion, where Akshay Kothari, Co-Founder and COO, describes Claude's "uniquely creative writing and summarization abilities" contributing to Notion AI.
- DuckDuckGo, for DuckAssist, described by Chief Business Officer Steve Fischer as "the first Instant Answer in our search results to use natural language technology to generate answers to search queries using Wikipedia and other related sources," with Anthropic "working closely with us to improve the quality of DuckAssist answers while also meeting our strict privacy requirements."
- Robin AI, a legal infrastructure business. CEO Richard Robinson states the firm has "been working with these types of technologies since 2019, and nothing has matched Claude's capabilities," and that it uses Claude "to evaluate particular parts of a contract, and to suggest new, alternative language that's more friendly to our customers." See Legal Services — AI Deployment.
- AssemblyAI, for its audio transcription and understanding APIs, quoted through founder and CEO Dylan Fox.
Provenance
Retrieved August 1, 2026 by direct HTTPS fetch from the canonical anthropic.com URL (HTTP 200), converted to text with a local extractor; Firecrawl was out of credits during this cycle. The page as served in 2026 carries the "Mar 14, 2023" dateline under the "Announcements" category. A pricing PDF linked from the post ("We're offering two versions of Claude today") resolves to an Anthropic CDN file whose filename carries a May 2023 date and was not retrieved; the post's own text was not relied on for pricing figures, and none are stated here.
Relationships
- related: Anthropic — the publisher.
- related: Claude 1 (Anthropic) — the models the post announces.
- depends-on: Constitutional AI — the training research the post's "helpful, honest, and harmless" formulation refers to.
- related: Constitutional AI: Harmlessness from AI Feedback — the research paper the deployed product follows.
- related: ChatGPT: Optimizing Language Models for Dialogue (OpenAI, November 30, 2022) — the launch four months earlier that the post's user testimonials compare against by name.
- related: Legal Services — AI Deployment — the Robin AI deployment is an early recorded legal-sector use.