Anthropic Unveils Claude Sonnet 4.5, Proclaiming It as Top Coding AI
On Tuesday, Anthropic launched its newest artificial intelligence model, dubbed Claude Sonnet 4.5, which it claims to be the “best coding model in the world.” This latest version is touted to improve coding abilities, better support computer use, and effectively cater to various business requirements, with notable strengths in fields like cybersecurity, finance, and research.
Unlike previous models, Claude Sonnet 4.5 can create fully functional applications ready for real-world use rather than just experimental prototypes. The company emphasizes that this model outperforms others in several critical coding benchmarks, including the SWE-Bench Verified test.
In a blog post, Anthropic highlighted that Claude Sonnet 4.5 marks a significant advance in the AI’s behavior, thanks to improved safety training. The model shows less inclination towards harmful behaviors, such as flattery or deception, and has enhanced security against prompt injection attacks, a major threat users face.
Anthropic has been on a roll, releasing new AI models almost every few months, with Claude Opus 4.1 launched just last August and Claude Sonnet 4 earlier in May.
The capabilities of Claude Sonnet 4.5 include running independently for up to 30 hours, focusing on complex tasks throughout that time. In comparison, its predecessor, Claude Opus 4, could only sustain operation for seven hours.
This enhanced model not only produces better quality code but also identifies areas for improvement and follows instructions more accurately.
Claude Sonnet 4.5 will be accessible through the Claude API and Claude chatbot, with pricing remaining unchanged at $3 per million input tokens and $15 per million output tokens.
Founded by former OpenAI researchers in 2021, Anthropic is a strong competitor in the AI landscape, now valued at approximately $183 billion. This valuation has surged since the introduction of its AI assistant, Claude, in March 2023.
The debut of Claude Sonnet 4.5 comes shortly after OpenAI’s release of GPT-5, which aims to enhance coding capabilities as well. Both companies are keen to attract enterprise clients increasingly turning to AI for software development, giving rise to the concept of “vibe coding.”
A newly coined term in the tech industry, “vibe coding” refers to the practice of using AI to assist in coding. This innovation allows even those without technical backgrounds to create functional applications and websites.
Headquartered in San Francisco, Anthropic boasts over 300,000 clients within just two years, a sharp rise from 1,000, as businesses from various sectors seek AI tools for development.
Following OpenAI’s example, Anthropic is expanding its global presence, setting up its first office in Asia in Tokyo and adding over 100 roles in Dublin and London, along with a research hub in Zurich. The company is also looking to fill leadership positions in several Asia-Pacific regions and continues to broaden its reach across Europe.
Anthropic recently reported a revenue run-rate exceeding $5 billion, significantly up from around $1 billion earlier this year.
