Anthropic announced its latest AI model, Claude Opus 4.6, on 5 February 2026, delivering significant advances in coding, longer autonomous task execution and professional-grade output creation. According to Scott White, Anthropic's head of product for enterprise, the industry is transitioning into a vibe working era in which AI increasingly handles complex professional tasks autonomously — while enterprise customers account for roughly 80% of Anthropic's business.
Claude Opus 4.6 is the first Opus-class model to feature a 1 million token context window in beta, alongside support for outputs of up to 128,000 tokens. On GDPval-AA — a benchmark measuring economically valuable knowledge work across finance, legal, and other domains — the model outperforms OpenAI's GPT-5.2 by approximately 144 Elo points, translating into a higher score roughly 70% of the time. The model also shows significant improvement in information retrieval across longer context windows: on MRCR v2, a needle-in-a-haystack retrieval test, Opus 4.6 scores 76% compared with just 18.5% for Sonnet 4.5. Anthropic additionally introduced agent teams in Claude Code, enabling multiple AI agents to work in parallel on a single project.
The model achieves industry-leading results on both Terminal-Bench 2.0 and Humanity's Last Exam, and Anthropic's automated behavioural audit found low rates of misaligned behaviour alongside the lowest rate of over-refusals among recent Claude models. Claude Code reached $1 billion in annualised run-rate revenue just six months after becoming generally available, and Anthropic's enterprise market share grew from near zero in early 2024 to approximately 40% of companies using it in production by January 2026, according to an Andreessen Horowitz survey.
Sources:
1.
Ashley Capoot
2.
3.
