Anthropic Unveils Claude Sonnet 4.5 Setting New Standards in AI Coding Abilities

Anthropic, the Amazon-backed startup valued at $183 billion, unveils Claude Sonnet 4.5, its latest artificial intelligence model, setting new benchmarks in coding and practical business applications.

The new model, according to Anthropic, excels in specialized fields such as cybersecurity, finance, and research. It is now available to all users, promising superior performance in generating high-quality code, identifying code improvements, and following instructions more reliably.

Claude Sonnet 4.5: A Smarter Colleague

"People are just noticing with this model, because it's just smarter and more of a colleague, that it's kind of fun to work with it when encountering problems and fixing them," says Jared Kaplan, Anthropic's co-founder and chief science officer.

According to industry benchmarks like SWE-bench Verified, Claude Sonnet 4.5 is hailed as the "best coding model in the world." The model can run autonomously for up to 30 hours, maintaining focus on complex, multistep tasks throughout that period, a significant improvement over its predecessor, Claude Opus 4.1, which could only run for seven hours.

Enhanced Safety and Reliability

Anthropic has also enhanced the safety features of Claude Sonnet 4.5. The company claims it has reduced concerning behaviors such as deception, power seeking, and sycophancy. Additionally, the model is more resistant to prompt injection attacks, making it less susceptible to malicious activities like exposing sensitive data.

"This is the biggest jump in safety that I think we've seen in the last probably year, year and a half," Kaplan adds.

Industry Context and Future Outlook

Anthropic, founded by former OpenAI researchers in 2021, competes directly with OpenAI, which recently launched GPT-5. The AI industry is witnessing rapid innovation, with both companies pushing the boundaries of what AI models can achieve.

Mike Krieger, Anthropic's chief product officer, notes that Claude Sonnet 4.5 will be the default for users, recommended for almost every use case. However, paid subscribers can still opt to use Opus, and users with specific workflows can choose an older generation of Sonnet if they prefer not to migrate immediately.

Looking ahead, Anthropic plans to release one or two more models before the end of the year, continuing its aggressive push in the AI arms race.

References

• Anthropic launches Claude Sonnet 4.5, its latest AI model that's 'more of a colleague'