Anthropic Unleashed: AI model Claude 3 has human abilities and is stronger than GPT-4

Anthropic claims that its new AI model Claude 3, which comes in three AI chatbot versions, beats GPT-4 in all areas

by Sededin Dedovic
SHARE
Anthropic Unleashed: AI model Claude 3 has human abilities and is stronger than GPT-4
© Noah Berger / Getty Images

Anthropic, a revolutionary AI venture founded by former employees of famous AI startup OpenAI, is making waves in the industry with its latest AI model, the Claude 3. This revolutionary model boasts superior performance compared to its predecessors, including the OpenAI GPT-4-based ChatGPT bot, as well as Google Gemini, on various key industry tests.

Of course, these are the claims of the developers of this AI chatbot, and they also made very bold claims about achieving "almost human abilities" in certain tasks. Anthropic pushes the boundaries of what AI can achieve and they have announced a real revolution, which they say you will see for yourself when you try this model and give it a chance.

Under the umbrella of the Claude 3 model, three different versions are presented: Haiku, Sonet and Opus. Powered by the Claude.ai chatbot, Sonnet is easily accessible for free with a simple email registration. On the other hand, Opus stands out as the most comprehensive and powerful Large Language Model (LLM), available through the "Claude Pro" subscription service for $20 per month.

In particular, Opus introduces multimodal capabilities, allowing it to seamlessly handle both text and image queries, a significant improvement over its predecessors. Anthropic has provided global availability for its Opus and Sonnet models, making them available through Claude.ai and the Claude API in 159 countries which represents great possibilities.

Although the Haiku version is yet to be released, the company has assured its availability soon. At the time of the development of such technologies and the sudden rise of artificial intelligence, it is very important to have strong competition and a large number of players on the market so that progress is constant.

Competition pushes programmers and developers to their limits, often resulting in revolutionary discoveries.

TechCrunch Senior Writer Devin Coldewey and Anthropic Co-Founder & CEO Dario Amodei speak onstage during TechCrunch Disrupt 2023© Kimberly White / Getty Images

An overarching feature of all Claude 3 models is their ability to engage in live user conversations and perform data extraction tasks quickly and accurately in real time.

Anthropic highlights the models' ability to deliver near-instant results and process more complex multi-step queries with increased accuracy. So, without any waiting at all, you can get a text of several thousand words or any other answer that interests you.

It is particularly impressive that such a speed of response to the task would be present even with the strongest model that is multimedia, and such models required a little more time until now. Opus, in particular, showed exceptional performance on tasks requiring cognitive power akin to college or graduate students, outperforming the OpenAI GPT-4 model by a remarkable 14.7 percent.

Furthermore, Opus outperforms the ChatGPT bot in various domains, including math, coding, reasoning, and knowledge-based tasks. This seems very impressive, but it's still better to wait a while until the comparisons are made by some of the most famous artificial intelligence analysts.

The advances introduced by Claude 3 go beyond performance improvements over previous versions. The Sonnet, for example, exhibits twice the speed of its predecessors, the Claude 2 and Claude 2.1, along with improved intelligence.

His expertise in tasks that require quick responses, such as information retrieval and sales automation, is noteworthy. Meanwhile, the Opus maintains speed comparable to the Claude 2 and 2.1 models while delivering a significantly increased level of intelligence.

The developers have warned that everyone will be surprised by the intelligence of Opus, so much so that it seems intimidating. The Haiku version, announced as the fastest and most cost-effective model on the market, boasts exceptional efficiency.

His ability to analyze complex research papers containing graphs and diagrams in less than three seconds speaks volumes for his agility and versatility. Moreover, Claude 3's skill in handling a variety of visual formats, including photos, charts, graphs and technical diagrams, offers invaluable assistance to businesses dealing with PDFs, diagrams or presentation slides.

Notably, models demonstrate a nuanced understanding of content requirements, reducing the likelihood of refusal to respond while being vigilant against potential harm. Anthropic explains that the development of the Claude AI is underpinned by ten secret foundational pillars of "righteousness".

Comprehensive training of the Claude 3 model includes both proprietary and publicly available datasets, using state-of-the-art hardware from Amazon Web Services and Google Cloud. Significant investments from tech giants such as Amazon and Google underscore the industry's recognition of Anthropic's groundbreaking contributions, with Amazon recently investing $4 billion in the venture.

Anthropic announced new investments in their company from some of the current leading companies, but they still pointed out that Google and Amazon are currently the biggest investors and that they deserve a lot for the success of this project.

The world of artificial intelligence continues to rapidly develop in all areas, and we will see if this is a breakthrough.

SHARE