There are no affiliate links in this article, and no AI company pays this site anything. Facts were checked on 17 October 2026 on Anthropic’s own pages; prices are US list prices before tax unless stated otherwise.
AI assistants series, part 3 of 6: 1. Grok · 2. ChatGPT · 3. Claude · 4. Meta AI · 5. Gemini · 6. The comparison
Claude is made by Anthropic. It’s the assistant we hear about most from people who write or code for a living, and the least from small business owners, who often haven’t heard of it. This is part three of six; Grok and ChatGPT came first, and the comparison closes the series.
How Claude has evolved: a timeline
Dates from Anthropic’s announcements and release notes. Checked on 17 October 2026.
2023: the first Claude
- 14 March 2023: Anthropic introduces Claude and a lighter, faster Claude Instant, with early partners including Notion, Quora and DuckDuckGo.
- 11 July 2023: Claude 2.
2024–2025: the three tiers
- 4 March 2024: the Claude 3 family introduces the naming still used today: Haiku, Sonnet and Opus, “in ascending order of capability”.
- 22 May 2025: Claude Opus 4 and Claude Sonnet 4.
2026: Fable, Cowork and the 5.5 family
- 1 September 2026: Claude Fable 5.1, plus Claude Mythos 5.1 for participants in Anthropic’s Project Glasswing only.
- 16 September 2026: Claude Cowork comes to every conversation, so you can hand Claude a bigger task, like a report, spreadsheet or presentation, without picking a mode first.
- 22 September 2026: Claude Opus 5.5, the first of the 5.5 family.
- 28 September 2026: Claude Sonnet 5.5.
- 7 October 2026: Claude Haiku 5.5.
The tiers in plain English
Anthropic names its models by size rather than just by version number, which helps when you’re choosing one:
- Haiku: the smallest and cheapest. Good for high-volume, simple jobs: tagging, short summaries, classifying form submissions.
- Sonnet: the middle tier, and the usual everyday choice.
- Opus: the larger model for long, complex work.
- Fable: sits above Opus on price. Anthropic says Opus 5.5 “performs at the level of Claude Fable 5.1 on most work” at a lower cost.
Mythos 5.1 isn’t generally available, so we’ve left it out of the table below.
Current models
Checked on 17 October 2026 on Anthropic’s models overview and platform release notes.
| Model | Released | API price per million tokens (input / output) |
|---|---|---|
| Fable 5.1 | 1 Sep 2026 | $10 / $50 |
| Opus 5.5 | 22 Sep 2026 | $4 / $20 |
| Sonnet 5.5 | 28 Sep 2026 | $2 / $10 |
| Haiku 5.5 | 7 Oct 2026 | from $0.10 / $0.50 |
All four have a 1 million-token context window, up to 128,000 tokens of output and a training cutoff of June 2026. Haiku 5.5 is the first Haiku with an “effort” setting, so you can trade speed for more thinking. Anthropic publishes benchmark results for these models; we haven’t reproduced them, so they’re not here.
Plans and prices
Checked on 17 October 2026 on Claude’s pricing page. US prices before tax; expect different figures in Spain or the UK once local tax applies.
| Plan | Price | Notes |
|---|---|---|
| Free | $0 | Chat on web, desktop and mobile |
| Pro | $20/month, or $17/month billed annually ($200 upfront) | More usage and models; Claude Code; Claude in Chrome and Microsoft 365 |
| Max | from $100/month | 5x or 20x Pro usage |
| Team | Standard $25/seat monthly ($20 annual); Premium $125 ($100 annual) | 2 to 150 seats |
| Enterprise | $20/seat/month billed annually, plus usage at API rates | 20-seat minimum |
Where Claude fits in web and agency work
Coding help. Claude Code is included from Pro upwards, and the 5.5 models are aimed squarely at what Anthropic calls “long-running agentic coding”. In WordPress terms, that’s things like refactoring an old theme’s functions file or writing a small plugin with tests. With a 1M-token window on every current model, you can give it a lot of a codebase at once. As with any assistant, review and test the output before it goes on a live site.
Content. Drafting and rewriting web copy, case studies and proposals. Cowork in every conversation (September 2026) means you can ask for a report or slide deck in the same chat as a quick question. As with any assistant, ask for British English (or Spanish from Spain) explicitly, and edit the draft before it goes out.
Research. Summarising long documents is where a big context window pays off: a 60-page tender, a client’s old brand guidelines, a long technical spec. Check anything factual against the source before you use it.
Business use. Team plans start at two seats, which suits a small agency. Claude for Microsoft 365 and Outlook is included on paid plans. Enterprise pricing is per seat plus usage at API rates, so costs depend on how much people use it.
Privacy, data and the EU
Claude’s pricing page shows model training as opt-out on consumer plans (Free, Pro and Max): your chats can be used for training unless you turn it off in settings. Team and Enterprise are commercial plans with their own terms. Check the current setting in your account’s privacy settings before you rely on it.
We found no EU or Spain-specific limits on the plans or models above in Anthropic’s release notes as of 17 October 2026. Prices are still shown in dollars, before tax.
The same rule for client work applies across this series: no client personal data or passwords in a consumer chat.
Pros and cons
Pros
- Clear model tiers (Haiku, Sonnet, Opus, Fable) that make cost and capability easy to match.
- A 1M-token context window on every current model.
- Claude Code included from the $20 Pro plan.
- Cowork turns bigger tasks (reports, spreadsheets, presentations) into part of a normal chat.
- Team plans from two seats, which suits small agencies.
- Opus 5.5 is cheaper than Opus 5 ($4/$20 against $5/$25).
Cons
- No low-cost tier between Free and Pro.
- Consumer chats are used for training unless you opt out.
- Enterprise adds usage at API rates on top of seats, which makes budgeting harder.
- Image and video generation aren’t its focus, compared with some rivals.
- Fewer consumer integrations than assistants built into big platforms (Google, Meta, X).
- The newest model (Mythos 5.1) is limited to a closed programme.
Who Claude suits
Claude suits people who write and code a lot and want a long context window for big documents or codebases. Pro is the obvious starting point for a solo consultant; Team works for a small agency that wants shared billing.
In short: a strong choice for people who write and code a lot, especially with long documents or codebases. Pro is the natural starting point for a solo consultant, and Team suits a small agency that wants shared billing. It’s less suited to anyone who mainly wants images or video.
Next in the series: part 4 of 6, Meta AI in 2026.
The AI assistants series
This is part 3 of 6.
- Grok in 2026
- ChatGPT in 2026
- Claude in 2026 (you are here)
- Meta AI in 2026
- Gemini in 2026
- Grok vs ChatGPT vs Claude vs Meta AI vs Gemini (the comparison)
Facts checked on 17 October 2026 on Anthropic’s own pages. AI products change monthly; we re-check models and prices before each update and note the date here.
