Claude: the precise one

Still one of the strongest models for code, agents and autonomous work.

Claude is Anthropic’s AI assistant, founded by former OpenAI staff with a focus on AI safety. With the current Opus/Sonnet generation and Claude Code as a CLI tool, Anthropic is a good address for software development and agentic workflows.

Versions compared (as of August 2026)

Free (Haiku)

Free

Quick, simple tasks with no sign-up hurdle.

  • Haiku model for everyday use

Popular

Pro (Sonnet)

from €17/month billed annually, €20/month when paid monthly

A balance of quality and speed for daily use.

  • Sonnet model, much higher usage limits than Free
  • Claude Code included

Max (Opus)

from €100/month

For heavy users, code, agents and complex tasks.

  • Unlimited access to the strongest model, Opus
  • Highest usage limits

Team

from €20/user/month

For small teams with a shared workspace.

  • Standard seat from €20, premium seat from €100/user
  • Central administration, privacy for companies

Enterprise

from €20/seat

Usage costs rise with model and task volume.

  • SSO, audit log, company-wide control
  • Custom contract terms

Strengths

  • Very large context window: whole codebases and books at once
  • Still one of the strongest models for code and agents
  • Claude Code: autonomous CLI tool for developers
  • Computer use: can control the desktop on its own
  • Comparatively few hallucinations
  • Projects feature for organising documents

Weaknesses

  • Web search has to be switched on in the chat (Anthropic’s guide); without it, Claude answers only from its training knowledge
  • No image generation of its own
  • Smaller ecosystem than ChatGPT
  • Sometimes noticeably cautious on sensitive topics

A comparison of all three major models is in the AI tools overview.

Pro tips

Use Claude Code

The CLI version is ideal for developers: right in the terminal, with file access and Git integration, no detour via the browser.

Use projects

Load all relevant documents into a project. Claude keeps track of the whole context instead of starting from zero with every new question.

Use XML tags

Claude understands structured prompts with <tags> noticeably better than plain running text, especially for longer tasks with several parts.

Common questions about Claude

Claude vs. ChatGPT – which is better?

For code, agents and long documents: Claude, thanks to its large context window. For versatility and image generation: ChatGPT. Many people use both, depending on the task.

What does the large context window mean in practice?

Claude can process very long inputs at once: whole contracts, books or codebases in one go, without forgetting the start of the conversation.

What is Claude Code?

A CLI tool that works autonomously in the terminal: reading and editing files, creating Git commits, running tests. Still one of the strongest options for developers.