Milestones
- Model Context Protocol open-sourced25 Nov 2024Complete.
- Claude Code research preview24 Feb 2025Complete.
- Claude Code generally available22 May 2025Complete.
- Claude Agent SDK released29 Sep 2025Complete.
- MCP donated to the Linux Foundation's Agentic AI FoundationDec 2025Complete.
Most important updates
- 9 Dec 2025
- 2 Dec 2025
- 20 Oct 2025
- 22 May 2025
- 25 Nov 2024
Current obstacles
- Safe autonomy on real machinesAgents with computer access can delete data or be tricked by hidden instructions in files and web pages. Permissions reduce the risk.
Physics limits
- Errors compound over long tasksIf each step succeeds 99% of the time, a 100-step task succeeds only about 37% of the time (0.99^100). Agents need near-perfect per-step reliability to work for hours.
- No program can prove code correct in generalRice's theorem: no algorithm can decide non-trivial properties of every program. Agents lean on tests, and code that passes weak tests can still be wrong.
- Summaries lose detailWhen a long session fills the context window, earlier work is compressed into summaries. Details dropped there are gone, so very long sessions drift.
How it works

An agent in your shell
Runs in the terminal, IDE, desktop and web with tools to read files, search, edit and run commands, working through a task step by step.
Asks before it acts
By default it asks before editing files or running commands. Users can pre-approve safe actions; web sessions run in a sandbox.
Project notes and context
A CLAUDE.md file holds project rules. When the context window fills, older work is summarised so long sessions can continue.
Plugs into other tools
Through the Model Context Protocol it connects to databases, issue trackers and browsers, and can hand side tasks to parallel subagents.
Papers & demos
- Nov 2024codeModel Context ProtocolOpen standard for plugging AI agents into tools and data, now used across the industry.
Update log
Tue 9 Dec 2025
- Minor: BlogPartnership
Tue 2 Dec 2025
- Major: BlogCommercial
Mon 20 Oct 2025
- Minor: BlogSoftware
Thu 22 May 2025
- Major: BlogSoftware
Mon 25 Nov 2024
- Major: BlogSoftware
About Anthropic
Anthropic
Claude models, Claude Code, AI safety
Anthropic is an AI safety company that makes the Claude models. Former OpenAI researchers founded it in 2021. It studies how models work inside, and Claude is widely used for coding and business agents.
- Raised $65B at a $965B valuation in May 2026, after a $30B round in February.
- Claude Code, launched in 2025, made Claude a leading model for AI coding agents.
- Trains on Google, Amazon and NVIDIA chips instead of relying on one supplier.
- Founded
- 20215 yrs
- Headquarters
- United States
- Status
- Private
- Valuation
- $965BprivateMay 2026
- Raised
- $131B6 rounds
- Last round
- Series H · $65BMay 2026$965B post
- Works in
- AIFrontier models
- Coverage
- 2 programs · 22 updateslatest 23h agochecked 25 Sep
- Lead investors
- Altimeter CapitalDragoneerGreenoaksSequoia CapitalGICCoatueMicrosoftNVIDIAICONIQFidelity
- Customers
- Salesforce
- Investors
- AmazonMicrosoftNVIDIA
- Suppliers
- People
- Dario AmodeiCEO and co-founderDaniela AmodeiPresident and co-founderJared KaplanChief Science Officer and co-founder

