Claude (Anthropic)
Claude is a family of large language models from the company Anthropic. It comprises the Haiku, Sonnet and Opus tiers and, with Claude Fable 5, a new flagship model placed above Opus. Claude is known for a very large context window, visible step-by-step reasoning and Constitutional AI, its safety-oriented training approach.
Also known as: Anthropic Claude, Claude AI, Claude models
What is Anthropic's Claude?
Claude is the language model family from the AI company Anthropic. Like other large language models, Claude understands and produces natural language and helps with tasks such as writing, analysis, research and programming. Anthropic places particular focus on the safety, reliability and traceability of the model's answers.
The current generation covers the Claude 4 family (Haiku 4.5, Sonnet 4.6, Opus 4.8) and the new, still more capable Claude Fable 5. Anthropic thus continues the series and combines strong reasoning with properties that make Claude particularly attractive for demanding, safety-sensitive use cases.
Which tiers are there: Haiku, Sonnet and Opus?
The Claude family is divided into three coordinated tiers. Haiku is the fastest and cheapest variant and suits simple, high-frequency tasks where speed and economy are to the fore. Sonnet is the balanced middle tier and offers a good relationship of capability, speed and cost for most everyday use cases.
Opus is the most capable tier and designed for particularly complex tasks, such as demanding reasoning and large programming projects. This tiering lets you choose the right model for the task and balance quality and cost deliberately.
Which Claude models are available now? (as of 2026)
The Claude family is built in performance tiers, from the fast entry-level model to the flagship. Currently available are: Claude Haiku 4.5 as the particularly fast, cost-efficient variant for simple, high-frequency tasks; Claude Sonnet 4.6 with the best relationship of speed and intelligence for most use cases; and Claude Opus 4.8 as the highly autonomous flagship of the Opus line for demanding knowledge work, complex coding and long-running agentic tasks.
At the top is Claude Fable 5 — currently the most capable generally available Claude model. It forms a new class above Opus and is built for particularly demanding reasoning and long, independent agent runs. It has a very large Context window and thinks along on every request, since the internal reasoning is always active. Closely related is Claude Mythos 5, which offers the same abilities but is reserved for selected organisations.
In practice this tiering means: Haiku, Sonnet and Opus cover the right balance of speed, cost and capability for different use cases, while Fable 5 comes in where the highest model quality and maximum depth of reasoning are needed. Which tier is optimal always depends on the specific use case.
What special capabilities does Claude have?
One outstanding feature of Claude is its very large context window. The model can thus process extensive documents, long conversations or whole code bases in one pass and take them into account as a whole. Claude additionally offers extended thinking, in which the model sets out its deliberations in visible step-by-step reasoning and so solves complex tasks in a more structured way.
Added to this is the ability Computer use, with which Claude can control software and browsers, for instance by entering input and operating interfaces. That allows tasks to be automated that previously required manual interaction with applications, and Claude can serve as a building block for AI agents.
What does constitutional AI mean?
Constitutional AI is the training approach developed by Anthropic that guides Claude towards safe, helpful answers. The model orients itself by a set of guiding principles, a kind of constitution, against which it assesses and improves its own answers. The aim is to reduce harmful or inappropriate output without limiting the model’s usefulness.
This approach underlines Anthropic's orientation towards trustworthy AI. For companies that value safety, compliance and traceable behaviour in particular, that is a relevant factor in choosing a model.
How can Claude be used in a company?
Claude can be used via an API into applications, platforms and workflows. Typical fields of use are processing large documents, knowledge research, demanding text work, programming support and building AI agents that handle multi-step tasks largely on their own.
Elisabit helps companies choose the right Claude tier, build the model sensibly into existing processes and account for safety, quality and economics. Claude's strengths, such as the large context window and extended thinking, can thus be put deliberately to work on concrete business requirements.
Frequently asked questions
What is Claude?
Claude is Anthropic's family of large language models. It comprises the Haiku, Sonnet and Opus tiers and is known for a large context window, visible reasoning and a safety-oriented training approach.
Which Claude models are available now?
Currently available are Claude Haiku 4.5 (fast and cheap), Claude Sonnet 4.6 (balanced between capability and speed) and Claude Opus 4.8 (the Opus line's flagship for complex reasoning and coding). Beyond these there is Claude Fable 5 as the most capable generally available model and Claude Mythos 5, which offers the same capabilities and is reserved for selected organisations.
Which is the most powerful Claude model?
The most capable generally available Claude model at present is Claude Fable 5. It forms a new performance class above the Opus line and is designed for particularly demanding reasoning and long, independent agent tasks. Claude Mythos 5 offers the same capabilities but is accessible only to selected organisations.
What is constitutional AI in Claude?
Constitutional AI is a training approach from Anthropic in which the model follows guiding principles to give safe, helpful answers. Claude assesses and improves its answers against these principles, to reduce harmful output.
What does computer use mean in Claude?
Computer use is Claude's ability to control software and browsers, for instance by entering input and operating interfaces. That allows tasks to be automated that would otherwise require manual interaction with applications, and Claude can serve as a building block for AI agents.
What is Claude good for in a company?
Through an API, Claude suits processing large documents, knowledge research, demanding text work, coding and building AI agents. The large context window and extended thinking make it particularly attractive for complex, safety-sensitive use cases.
Related terms
The most capable tier of Anthropic's Claude family, for complex reasoning and coding.
An LLM is an AI language model that understands and produces text by predicting the most likely next word.
Generative AI independently creates new content — text, images, audio or code — based on patterns it has learned.
GPT-5 is OpenAI's current frontier model generation, with strong reasoning and multimodal capabilities.
An AI agent's ability to operate a computer the way a person does — via screen, mouse and keyboard.
An AI system that pursues goals on its own: perceiving, planning, using tools and acting across several steps.
Put AI to work for your business?
We help you integrate artificial intelligence into your processes, your marketing and your website — strategically and securely.
Your contact
Stefan
I look forward to hearing about your project and finding the best solution together.
