Share with others:
Europe, here is Claude, the great rival of ChatGPT
It's true that the company OpenAI had the main say in the advent of generative artificial intelligence with ChatGPT, but besides it there are other players on the market that have quickly become synonymous with the hottest technology of the moment. When Anthropic, a startup heavily backed by Amazon, launched its first major language model, Claude 3, in March of this year, we users in Europe were disappointed that it was not yet available for use here.
On May 14, one of the main rivals of ChatGPT also came to us (the API was already available for developers a few months earlier).
The main thing we can say after less than an hour of testing is that it is fast. Really fast. First impressions are that Claude is more thorough in many responses, even when compared to the paid version of ChatGPT. He also seems more… human? And it's not just because Claude starts the first contact by asking the user by name and saying that he's glad to meet you.
While it's still fairly easy to tell when a text is written by ChatGPT, we're not entirely sure if the same can be said for Cloud. This fact is both impressive and alarming.
The current "Claude family" consists of three versions - Opus, Sonnet and Haiku. Opus is Anthropic's "most intelligent" model. According to developer data, it outperforms its competitors in most benchmarks for artificially intelligent systems.
Claude is currently available through the Claude.ai website, an iOS app, and a special group plan for business users. As free users, we initially have access to the Sonnet version, but it has a daily usage limit. Opus is available with a Pro subscription that costs 18 euros plus tax.
Claude claims that he "understands well and is fluent in French, German, Italian, Spanish and other European languages". Regardless, when prompted, he informs that he does not speak Slovenian fluently. Nevertheless, he writes from time to time - in Slovenian - so that he can offer us simpler answers in Slovenian, but it would be easier to discuss more complex topics in English. This illustrates one of the reasons why training large language models on data from less widespread languages is extremely important for all regions.
Claude is not multimodal, which means it is not capable of generating images like, say, OpenAI's DALL-E, which is included in the paid version of ChatGPT, among others. But Claude can process and analyze visual inputs.
On the same day that Anthropic released Cloud, OpenAI launched GPT-4o. Both companies are financially supported by technology giants (the first is Amazon and Google, the second is Microsoft). In addition to competing in the classical areas of large language models, Anthropic positions itself as a company committed to the development of secure artificial intelligence solutions.
