Claude is Anthropic’s AI assistant. You can use it in a browser, in desktop and mobile apps, and through connected tools such as Microsoft 365, Chrome and Slack. Anthropic now describes it less as a chatbot and more as an assistant that takes on tasks and produces files, with Claude Code for programming and Cowork for other multi-step work.
This review sets out what Anthropic documents about Claude’s features, how its individual and team plans are priced and limited, what its data-use choices mean, and who it suits. It is written for people deciding which assistant to pay for, including small teams without a technical specialist.
How we researched this: we did not run hands-on tests for this article. It is based on the vendor’s own documentation and public sources, checked on 11 November 2026. Features and prices change, so confirm details on the vendor’s site before you buy.
Key takeaways
- Anthropic’s pricing page lists Free, Pro ($20 monthly), Max ($100), Ultra ($200), Team and Enterprise plans. Higher individual tiers are defined mostly by usage multiples of Pro.
- Usage is limited per five-hour window and per week, shared across models, with a separate weekly limit for the Fable model family. Anthropic says it does not guarantee uninterrupted service.
- Claude’s documented work features include Projects, Artifacts, Claude Code, Cowork and connectors. Cowork is not on the Free plan.
- For Free, Pro and Max accounts, Anthropic’s August 2025 terms update gives a choice about using chats to improve models. Opting in means five-year retention, otherwise 30 days. Team, Enterprise and API use fall under commercial terms and were not affected.
- Anthropic itself warns that Cowork carries unique risks because it is agentic and has internet access, so people should review actions on sensitive files.
What it is
According to Anthropic’s product overview, Claude is an AI assistant for work. You point it at files and tools, describe the outcome in plain language, and it does the work, including editable files such as decks, spreadsheets and prototypes. Short questions get quick replies, and longer tasks can keep running after you close your laptop. The page says Cowork and chat are now integrated within Claude.ai.
Claude is built on large language models, which can state wrong things with confidence. If that term is new, read what is a large language model.
Models
Anthropic’s pricing page lists current models and their API prices per million tokens (input and output): Fable 5.2 at $8 and $40, Opus 5.5 at $4 and $20, Sonnet 5.5 at $2 and $10, and Haiku 5.5 at $0.10 and $0.50. These are API prices, which are separate from the monthly plans. Legacy models are also listed. Which models each plan can reach is described below.
Who it is for
- Individuals and freelancers who want an assistant for writing, analysis, research and file-based work, on a monthly plan.
- Heavy users, such as people running long sessions or coding every day, who may look at Max or Ultra.
- Small and mid-sized teams, since Team is listed for 2 to 150 seats with central billing and single sign-on.
- Larger organisations, who can look at Enterprise for admin spend limits, SCIM, audit logs and a compliance API.
If you only need occasional answers, the Free plan may be enough. If your work is entirely inside another vendor’s workspace, compare that vendor’s own assistant first.
Key features (documented)
| Feature | What Anthropic documents |
|---|---|
| Chat | Quick answers and longer tasks that can continue in the background |
| Projects | Keep context, instructions and files organised by project |
| Artifacts | Shareable designs, slides and documents you can edit yourself or revise by asking Claude |
| Claude Code | Coding work under the same subscription on Max, with a separate product page |
| Cowork | Hand off work that runs in parallel, on a schedule or from your phone, with editable output |
| Integrations | Connectors to apps, Claude in Chrome, Claude for Microsoft 365, Slack, Google Workspace, custom connectors and MCPs |
| Other | Skills, plugins, memory, file handling across many formats, and voice |
Cowork in more detail
Anthropic’s help centre says Cowork brings the approach of Claude Code to knowledge work outside coding. You describe an outcome, and Claude plans and carries out multi-step tasks such as organising files, producing spreadsheets and slides, or synthesising research. Sessions run in the cloud. The desktop app is available on all paid plans, and web and mobile are listed for Pro, Max, Ultra and Team (beta), with Enterprise when an owner enables it. Free plans do not include Cowork.
Documented limitations include that sessions cannot be shared with others (individual artifacts can), local file and browser access need the desktop app open, and multi-step tasks use substantially more usage than simple questions. For a plain-English explanation of this type of tool, see AI agents explained for non-technical readers.
Pricing model
Checked on 11 November 2026, US dollars. Anthropic states that prices and plans are subject to change. Mobile pricing may differ.
| Plan | Monthly | Annual | Usage and notes |
|---|---|---|---|
| Free | $0 | n/a | Limited usage per five-hour and weekly window; up to 5 Projects; Sonnet and Haiku |
| Pro | $20 | $17 per month ($200 up front) | 1x usage; Claude Code, Projects, Opus, Claude in Chrome, Microsoft 365; usage credits |
| Max | $100 | n/a (monthly only) | 5x Pro usage; $100 monthly API credits; priority access at high traffic |
| Ultra | $200 | n/a (monthly only) | 10x Pro weekly usage; no five-hour limit for verified accounts; $200 monthly API credits |
| Team | $25 standard seat, $125 premium seat | $20 and $100 | 2 to 150 seats; premium seats get 5x standard usage; up to $500 monthly pooled API credits; SSO and central billing |
| Enterprise | Not listed | $20 per seat, 20-seat minimum, plus usage at API rates | Admin spend limits, SCIM, audit logs, compliance API, HIPAA-ready offering, Claude Security (beta) |
How limits work
- Anthropic describes a five-hour session limit and a weekly limit shared across all models, plus a separate weekly limit for Fable.
- The weekly reset happens at a fixed time assigned to your account.
- Max and Ultra include Fable at up to 50 percent of weekly limits. Pro offers Fable through usage credits.
- Anthropic may add other caps and does not guarantee uninterrupted service.
The pricing page and a help page describe the plans slightly differently in places (for example, the help page we read did not list prices for Pro and Ultra), so rely on the pricing page and confirm. See also how to read an AI tool’s pricing page.
Strengths
- Task and file focus. Projects, Artifacts and Cowork are aimed at producing finished documents, not only chat replies.
- Integrations. Connectors, Microsoft 365, Chrome and Slack support, plus MCP, are documented.
- Permission controls. Anthropic says permission settings can make Claude show its plan and wait for approval, and that it asks before permanently deleting files by default.
- Visible edits. In Word and Excel, changes appear as tracked edits or highlighted cells, per the overview page.
- Admin features for organisations. Spend limits, tool permissions by department and SIEM streaming through OpenTelemetry are listed for Team and Enterprise.
Limits and cautions
| Caution | Detail |
|---|---|
| Usage caps | Five-hour and weekly limits apply, so heavy users may hit them. Higher tiers raise limits but cost more. |
| Agentic risk | Anthropic says Cowork has unique risks due to its agentic nature and internet access, and that in auto mode “no defense is perfect”. |
| Skip mode | Skip mode runs without checks, and Anthropic advises using it only when you fully trust every action, connector and file involved. |
| Network settings | Anthropic notes network egress settings do not cover web fetch, web search, MCPs or Claude in Chrome. |
| Free tier gaps | No Cowork on Free, and a smaller set of models. |
| Wrong answers | Like any language model, Claude can make mistakes. Use how to check AI output for mistakes. |
Privacy and data
Anthropic published an update to its consumer terms and privacy policy on 28 August 2025. It affects Free, Pro and Max accounts, and Claude Code sessions from those accounts.
| Question | What Anthropic says |
|---|---|
| Training on chats | Users can choose whether their conversations are used to improve Claude. The choice can be changed in Privacy Settings. |
| Retention if you allow it | Five years, including feedback you submit. |
| Retention if you do not | The existing 30-day period applies. |
| Already-trained data | Turning the setting off excludes new chats from future training runs. It does not affect data already used. |
| Not affected | Services under commercial terms: Team, Enterprise, Claude Gov, Claude for Education, the API, and access via Amazon Bedrock or Google Cloud’s Vertex AI. |
The terms page was dated 2025, so check Anthropic’s current privacy policy for any later changes. We could not confirm from the pages we read what the setting shows by default for new accounts today, so open Privacy Settings after you sign up. For team policy, read how to protect company data when using AI tools.
Alternatives
- ChatGPT for a broad consumer assistant with its own tiers. See the ChatGPT review.
- Gemini if you live in Google’s apps. See the Gemini review.
- Notion AI if your knowledge sits in Notion. See the Notion AI review.
Who should and should not use it
Best for you if
- You want an assistant that produces files and can work on longer tasks, and you will review its work.
- You want a team plan with central billing and SSO.
- You are comfortable managing usage limits across the week.
Look elsewhere if
- You need a flat, unlimited-feel plan. Limits are explicit.
- Your data cannot be used for model improvement and you will not manage the setting. Use a commercial-terms plan instead.
- You want agentic features but cannot review actions on sensitive material.
Questions to ask before buying
- How much usage do I need, and which tier matches it?
- Do I need Fable access, and is it included in my plan or billed through credits?
- Is the model-improvement setting off for every user?
- Are connectors approved by IT, and who can add MCPs?
- Which permission mode will we use for Cowork?
- Does annual billing exist for the tier I want? On the pricing page we read, Max and Ultra were monthly only.
Run a short trial following how to test an AI tool before you pay for it.
Frequently asked questions
How much does Claude cost?
Anthropic’s pricing page lists Pro at $20 per month, Max at $100 and Ultra at $200, with Team seats at $25 or $125. Check the live page.
Is there a free version?
Yes, with limited usage and up to five Projects, using Sonnet and Haiku models, according to the pricing page.
Does Anthropic train on my chats?
For Free, Pro and Max, Anthropic’s 2025 terms update says users can choose whether chats are used to improve Claude. Commercial plans were not affected.
Does Claude Code cost extra?
The pricing and help pages say Claude Code is included under Pro and Max subscriptions. Check usage limits, which apply across products.
What is Cowork?
Anthropic’s agent-style feature for multi-step, non-coding tasks. It is not on the Free plan.
Which assistant should I pick?
Use how to choose an AI tool and test with your own tasks.
Conclusion
Claude’s documented strengths are file-oriented work, integrations and permission controls. The costs are explicit usage limits, tiers defined by multiples of Pro, and agent features that Anthropic itself says need supervision. For privacy, check the model-improvement setting on personal plans.
Compare it fairly with how to compare AI tools side by side fairly, and confirm plan details on Anthropic’s pricing page before you buy.
Sources
- Claude pricing, accessed 11 November 2026
- What is the Max plan (Claude Help Center), accessed 11 November 2026
- Claude product overview, accessed 11 November 2026
- Get started with Claude Cowork (Claude Help Center), accessed 11 November 2026
- Updates to our consumer terms (Anthropic), 28 August 2025, accessed 11 November 2026
