How we review: this assessment is based on the product’s official documentation, current pricing and published independent evaluations; it is not a hands-on test. Read our review methodology.
Anthropic shipped a full refresh of its model line-up in September and October 2026: Claude Opus 5.5 on 22 September, Sonnet 5.5 on 28 September and Haiku 5.5 on 7 October, according to the company’s newsroom. Around those models, the Claude apps have grown from a chat window into a workspace with projects, long-running tasks, connectors and a coding agent. This review assesses the product as Anthropic documents it today.
What it is
Claude is Anthropic’s AI assistant, available on the web, on iOS and Android, and as desktop apps for macOS and Windows. It is built on the same models Anthropic sells through its API and through Amazon Web Services, Google Cloud and Microsoft Azure. Our directory entry on Claude gives a short overview; here we focus on value and fit.
Key features
- Models. The current family is Haiku (fast and small), Sonnet (balanced) and Opus (most capable in the main line), with a separate premium model, Fable, offered on some plans. Anthropic says Opus 5.5 performs at about the level of Fable 5.1 on most work while costing roughly 40% less to run than Opus 5 on typical workloads. Those are vendor claims, and Anthropic’s own benchmark table shows Opus 5.5 ahead on some tests and behind competitors on others.
- Projects and Artifacts. Projects keep instructions, files and context together; Artifacts are shareable documents, designs and slides you can edit directly or ask Claude to revise.
- Tasks. The product overview describes longer tasks that keep running after you close your laptop, can be started from a phone, and can run in parallel or on a schedule.
- Connectors, skills and plugins. Claude connects to other apps and data sources, including custom connectors built on the Model Context Protocol. Teams can record and share reusable skills.
- Claude in Chrome and Microsoft 365. On paid plans, Claude can act inside the browser and inside Microsoft 365 apps; in Excel, Anthropic says it flags every cell it changes and adds notes explaining why.
- Claude Code. Anthropic’s agentic coding tool is included from the Pro plan upwards.
- Controls. Users can require Claude to show a plan and wait for approval, and restrict which folders and tools it may access.
Pricing
At the time of writing (October 2026), Anthropic’s official pricing page lists (US dollars, excluding tax):
| Plan | Price | Highlights |
|---|---|---|
| Free | $0 | Sonnet and Haiku, web search, file creation, memory, connectors, Artifacts, up to 5 projects |
| Pro | $20/month, or $17/month billed annually ($200 up front) | Opus, Claude Code, unlimited projects, Claude in Chrome, Microsoft 365, scheduled tasks; Fable via usage credits |
| Max | From $100/month | 5x or 20x Pro usage, priority access, $100–$200/month in API credits |
| Team Standard | $20/seat/month annual, $25 monthly | 2–150 users, SSO, admin controls |
| Team Premium | $100/seat/month annual, $125 monthly | 5x a Standard seat, includes Fable |
| Enterprise | $20/seat/month billed annually (20-seat minimum) plus usage at API rates | SCIM, audit logs, compliance API, HIPAA-ready option |
For developers, Opus 5.5 is priced at $4 per million input tokens and $20 per million output tokens on the API, per Anthropic’s launch post.
Strengths
Claude’s free tier is unusually generous in features. Memory, connectors, Artifacts and file creation are available without paying, which makes it easy to judge fit before committing.
The Pro plan is strong value for anyone who codes. Claude Code alone is a reason many developers subscribe, and it comes bundled with the general assistant, browser agent and Microsoft 365 integration at the same $20 price as most rivals’ standard tiers. Our guide to using AI coding assistants effectively explains how to get the most from agents like this.
Anthropic also publishes a detailed safety and behaviour story with each release — for example, it says Opus 5.5 scored best among its tested models on its automated behavioural audit and resisted prompt injection better. Organizations with strict risk policies tend to value that documentation, and the Team and Enterprise tiers add SSO, SCIM, audit logs and a HIPAA-ready option.
Limitations
Every plan has usage limits. Anthropic raised the five-hour limits with the Opus 5.5 launch, but people who run long agentic coding sessions can still hit them, which is why Max exists.
The model line-up is getting complicated. Fable is available through usage credits on Pro, at “50% of weekly limits” on Max and Team Premium, and fully on Enterprise. Understanding which model you get, and at what cost, takes reading.
Max is listed only as “from $100” with a choice of 5x or 20x usage; the pricing page does not itemize the higher tier. Several features — Claude Dashboards, Claude Security and Claude in Slack — are labelled beta.
Who it’s for
- Writers, analysts and researchers who work with long documents and want polished prose.
- Developers who want a capable coding agent bundled with a general assistant.
- Teams on Microsoft 365 who want an assistant that works inside Excel and other Office apps.
- Regulated organizations that need audit logs, retention controls and clear safety documentation.
Alternatives
- ChatGPT — broader consumer feature set and a cheaper $8 tier.
- GitHub Copilot — if you mainly want coding help inside your editor and GitHub.
- Perplexity — better suited to cited, search-first research.
Our ChatGPT vs Claude vs Gemini guide explains how to compare these assistants on your own tasks.
Verdict
Claude earns 4.5 out of 5. The documented feature set is excellent for writing, analysis and coding, the free tier is generous, and Pro bundles Claude Code and Office integration at a mainstream price. Strong admin and compliance options make it easy to recommend to organizations. We hold back half a point because usage limits remain a real constraint for heavy users and the growing mix of models, credits and beta features makes the plans harder to understand than they should be.
