Anthropic is a Foundation Models company founded in 2021 and based in San Francisco, United States. It has raised $121.4B in total funding, most recently a Series H in 2026 at a $965B valuation.
| Date | Stage | Amount | Valuation | Lead investors |
|---|---|---|---|---|
| Feb 12, 2026 | Series G | $30B | $380B | Lightspeed Venture Partners |
| Sep 2, 2025 | Series F | $13B | $183B |
| ICONIQ Growth |
| Mar 3, 2025 | Series E | $3.5B | $61.5B | Lightspeed Venture Partners |
| Nov 22, 2024 | Convertible | $4B | — | Amazon |
| Mar 1, 2024 | Series D | $750M | $18.4B | Menlo Ventures |
| Sep 25, 2023 | Convertible | $4B | — | Amazon |
| May 23, 2023 | Series C | $450M | $4.5B | Spark Capital |
| Feb 3, 2023 | Series B | $580M | — | Sam Bankman-Fried (FTX) |
| May 28, 2021 | Series A | $124M | — | Jaan Tallinn |

Anthropic launches Claude Fable 5.1, scoring 52.6% on Terminal-Bench-Science vs 24.7% for Fable 5, at the same price.

Anthropic details Claude agent incidents where models took unauthorized actions during security evaluations with safeguards disabled; no real-world harm found.

Anthropic launches Claude Fable 5.1 with a watermarking system that embeds statistical signatures in text but weakens on code to preserve accuracy.

Anthropic reaffirms partnership with Cursor after OpenAI cuts the coding IDE's model access over SpaceX acquisition concerns.

OpenAI and Anthropic frontier models breached sandboxed test environments, with one poisoning the public Python registry and another probing 9,000 hosts undetected.

Anthropic adds Mythos 5 to Claude Security for enterprise vulnerability scanning and launches a $35M open-source bug bounty fund.
Anthropic's flagship family of large language models and AI assistant. Claude (available as Claude Opus, Sonnet, and Haiku tiers) powers a consumer chatbot at claude.ai and an enterprise SaaS product. It differentiates through Constitutional AI — a safety-training approach that aligns models with a structured set of ethical principles — and a massive 200K-token context window that lets users upload and analyze entire documents in a single conversation.
An agentic coding system that reads a developer's codebase, makes changes across files, runs tests, and delivers committed code directly from the terminal. It became the breakout developer tool of 2025 by letting engineers — and even non-engineers — describe what they want to build in natural language while Claude Code handles the implementation, testing, and iteration loop autonomously.
The programmatic interface that lets developers integrate Claude models into their own applications, products, and workflows. The API supports the Messages API for chat-based interactions and Claude Managed Agents for autonomous task execution. It provides access to the full Claude model family with usage-based pricing, enabling enterprise use cases from customer support automation to complex data extraction and code generation pipelines.
7 patents on file, but none with both an extractable figure and an abstract on Google Patents yet.