Anthropic, an American startup founded by former OpenAI employees and developing tools based on artificial intelligence, announced that it has launched its new chatbot Claude 2. Guard.
Claude 2, like the previous version of Claude 1.3, can search for information about documents, write text and answer questions. Unlike ChatGPT, it can note around 75,000 words, not 3,000.
According to Startup, the Claude 1.3 model performs better on questions in the professional medical license exam with multiple choices. It also performs better when answering the same types of questions in the test for lawyers – 73% versus 76.5%.
The company recommends thinking of “Claude 2 as a friendly, enthusiastic colleague or personal assistant who can be informed in natural language.”
Anthropic claims that the updated neural network writes code better, solves math problems, generates longer text, and formats responses correctly in JSON, XML, and YAML more often. But it cannot search for information like GPT-4 on the internet. The AI bot is already available for US or UK residents on the Anthropic website.
Formerly former Google engineer saidthat the company is trying to approach the latest versions of AI in a “safe and responsible” way.