Anthrope yesterday (May 22) launched Claude Opus 4 and Claude Sonnet 4 – the models claim that they establish new standards for coding, advanced reasoning and AI.
Claude Sonnet 4 replaces Claude Sonnet 3.7, with better coding and reasoning while responding more accurately to the instructions, says Anthropic, which described Claude Opus 4 as his most powerful model so far and “the best coding model in the world.”
“Built for sustained performance in complex and long -lasting tasks, it can maintain an effort focused on thousands of steps,” said the company on Opus 4.
The two new models are hybrid and sacrifice two instantaneous responses of extended modes or thought for deeper reasoning.
“These models advance the strategies of the AI of our clients in all areas,” Anthrope said in his statement. “Opus 4 pushes the limits in coding, research, writing and scientific discovery, while Sonnet 4 provides border performance to daily use cases such as an instant update of Sonnet 3.7”.
Anthrope, who recently announced an important European expansion, including jobs in Dublin, also launched a complete game or updates to Claude.
“Claude 4 models raise the ribbon for experienced engineers, which facilitates writing, edition and purification code while promoting the tools that developers love, and open complete beginners, from Solpreneurs to sales leaders, putting creative power in the Howsim, that.
Anthrope was founded by brothers and students of Ai Open Dario and Daniela Amodei. Even a private company was based on the principles of the responsible AI, although many commentators have questioned the ability of any AI player to be truly ethical.
In recent years he has received funds from people like Amazon and Google, attracting purists online.
Anthrope launched the new models with a lot of criticism of illustrious users who sprouted about their effectiveness, but one caught my eye, a review of Claude Sonnet 4 of Tao Zhang, co -founder of the main AI player.
Manus launched a subscription offer of $ 39/month in March, putting it in direct competition with the tastes of Chatgpt and, of course, Claude, although Manus Itelf admits that it is still in the test phase. Investors reported in Manus include Benchmark, Tencent Holdings, HSG (previously Sequoia) and Zhenfund.
“Claude 4’s ability to follow complex instructions and several steps and work through problems with a clear reasoning of the thinking chain is remarkable. The aesthetics of artifacts is really an exception: I have seen something like that.”
As for Claude Opus 4, Yusuke Kaji, general manager, AI, in Rakute, was very impressed: “Opus 4 sacrifices a truly advanced reasoning for coding,” he said. “When our team displayed Opus 4 in a complex open source project, it is coded to autonomous for almost seven hours, a great leap in the abilities of AI that left the astonished team.”
De facto praise. As Claude user, I have no illusions that Anthrope is perfect, but I had a great time for his ethical aspirations, I am excited to break it when trying it. Not to inform, I hurry to add: we have a strict policy of here in Siliconrepublic.com when it comes to journalism. I will inform once you run the new model through its steps for more mundane tasks.
Do not miss the knowledge you need to succeed. Register in the daily letter, the Silicon Republic Summary of need for SCI-TECH.
