Late at night, a big move was secretly made, and suddenly, I had a colleague who could work for 30 hours straight! Who can outcompete that?

Late at night, a big move was secretly made, and suddenly, I had a colleague who could work for 30 hours straight! Who can outcompete that?

Just past midnight last night, AI company Anthropic officially launched the next-generation model, Claude Sonnet 4.5. The company directly states that it is the best coding model in the world, the most powerful tool for building complex agents, and the best model for using computers.

The Claude Sonnet 4.5 API is now available on DeployAI, with the pricing matching the official website. Users in China can directly access it by clicking the link below!

Claude Sonnet 4.5 entry or simply copy and paste into your browser (https://deployai365.com/ApiExplorer)

Claude Sonnet 4.5 ranks at the top with a 77.2% accuracy in the SWE-bench Verified benchmark, an improvement of nearly 20 percentage points over its predecessor. It excels in engineering implementation. During testing, it autonomously ran for 30 hours and generated 11,000 lines of code, fully developing an enterprise application similar to 'DingTalk', covering the entire process from database configuration, domain registration, to compliance auditing, becoming the first AI model to achieve 'production-level' delivery.

At the same time, Claude Sonnet 4.5 demonstrates exceptional domain knowledge and reasoning ability in fields such as finance, law, medicine, and STEM.

Claude Sonnet 4.5 no longer categorizes for pleasing humans, reducing emotional considerations (less flattery, deception, power pursuit, encouragement, etc.), but continues the improvements in Claude's capabilities and safety measures. It has made significant progress in agent modeling and computer usage abilities, especially in detecting network attacks.

However, the pricing for Claude Sonnet 4.5 remains at $3.45 per million tokens (making it the most expensive coding AI model).

As for whether it can sustain 30 hours of continuous work and maintain its impressive accuracy, I'm still testing it. If you want to know the results, make sure to like and follow, and I'll update you with the latest findings!

Claude Sonnet 4.5 entry or simply copy and paste into your browser (https://deployai365.com/ApiExplorer)