Anthropic Claude 4 Review: Creative Genius Trapped by Old Limitations
San Francisco-based Anthropic has launched the fourth generation of its Claude AI models, positioning itself as an outlier among major AI companies. While competitors like Google and OpenAI push boundaries with expanded context windows and multimodal capabilities, Anthropic maintains a 200,000-token limit and text-only approach.
The timing of Claude 4’s release appears strategic, coinciding with Google’s Gemini announcement and OpenAI’s new Codex-based coding agent. Anthropic’s hybrid models—shifting between reasoning and non-reasoning modes—offer a glimpse of what OpenAI might deliver with GPT-5.