Anthropic's latest AI model can code autonomously for 30 hours
What's the story
Anthropic has unveiled its latest artificial intelligence (AI) model, the Claude Sonnet 4.5. The company claims this new model is the "best coding model in the world." It excels at programming, using computers, and fulfilling real-world business requirements. The advanced system also shines in specialized areas like cybersecurity, finance, and research.
Enhanced capabilities
The new model can create production-ready applications
The Claude Sonnet 4.5 can create "production-ready" applications, a major improvement over its predecessors. The model also provides top-notch performance on various coding benchmarks, including SWE-Bench Verified. This new system is more aligned than ever before, with improved capabilities and extensive safety training to reduce problematic behaviors such as sycophancy, deception, power-seeking tendencies, and promoting delusional thinking.
Autonomous operation
It can work independently for up to 30 hours
The Claude Sonnet 4.5 can work independently for up to 30 hours, maintaining focus on complex, multistep tasks throughout that period. This is a major improvement over its predecessor, the Claude Sonnet 4, which could only run autonomously for seven hours. The new model will be available to all users through the Claude API and chatbot at the same pricing as its predecessor: $3 per million input tokens and $15 per million output tokens.
Market growth
Anthropic is opening its first Asia office in Tokyo
Founded in 2021 by former OpenAI researchers, Anthropic has been aggressive with its AI model launches. The company is currently valued at $183 billion. In just two years, it has grown from 1,000 customers to over 300,000. Following OpenAI's lead, Anthropic is opening its first Asia office in Tokyo and expanding across Europe with over 100 new positions in Dublin and London.