Anthropic’s Claude 3 model outperforms GPT-4 and Gemini Ultra in many tests

Anthropic announced on Monday a new family of AI models, collectively called the Claude 3 model family. As is commonly done, the company released three different sizes of models, each with a varying balance of intelligence, speed, and cost.

The largest of the new models, called “Opus,” outperforms both OpenAI’s and Google’s most advanced models, GPT-4 and Gemini Ultra, respectively, on tests measuring undergraduate level expert knowledge (MMLU), graduate level expert reasoning (GPQA) as well as basic mathematics (GSM8k), Anthropic says.

The middle child in the family, Claude 3 “Sonnet,” is twice as fast as Anthropic’s previous best model, Claude 2.1, and with higher intelligence. Anthropic says Sonnet excels at intelligent tasks demanding rapid responses, like knowledge retrieval or sales automation.

The smallest model, called “Haiku,” beats other comparably sized models in performance, speed and cost, the company says. It can read a dense research paper of roughly 7,500 words with charts and graphs in less than three seconds.

All three models can process visual imagery, which enables them to understand uploaded documents, analyze web interfaces, and generate image catalog metadata. Anthropic says that for many of its enterprise customers, up to half of their knowledgebases consist of documents in image formats such as PDFs, flowcharts, or slides.

The Opus and Sonnet models are available today, while the Haiku model will be available soon.

https://www.fastcompany.com/91046925/anthropic-claude-3-models?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss

Vytvořeno 2y | 4. 3. 2024 21:10:09


Chcete-li přidat komentář, přihlaste se

Ostatní příspěvky v této skupině

This viral grocery hack will help you save money and reduce waste

If you dread the weekly grocery shop, or get sidetracked by fun snacks only to end up with no real meals, this might be the hack for you.

The 5-4-3-2-1 method gives shoppers like you a s

31. 8. 2025 13:10:02 | Fast company - tech
Do Trump’s tariffs mean you’ll pay more for the iPhone 17 next month?

If 2025 is the year of anything, it is the year of the tariff. Ever since President Trump unleashed his

30. 8. 2025 11:30:07 | Fast company - tech
This simple free service makes sharing PDFs painless

Look, I’m not gonna lie to ya’: I’ve got a bit of a love-hate relationship with PDFs. And, more often than not, it veers mostly toward the “hate” side of that spectrum.

Don’t get m

30. 8. 2025 11:30:04 | Fast company - tech
Palantir is mapping government data. What it means for governance

When the U.S. government signs contracts with private technology companies, the fine print rarely reaches the public. Palantir Technologies, however, has at

30. 8. 2025 9:10:09 | Fast company - tech
‘The New York Times’ paywalled the Mini Crossword and the internet is in shambles

Bad news for morning routines everywhere: The New York Times has put its Mini Crossword behind a paywall.

On Tuesday, instead of their usual puzzle, players were met with a paywall. The

29. 8. 2025 19:20:05 | Fast company - tech
Chinese tech giant Alibaba aims to fill Nvidia void with its new AI chip

China’s Alibaba has developed a new chip that is more versatile than its older chips and is meant to serve a broader range of

29. 8. 2025 16:50:06 | Fast company - tech
How Japan is using AI to prepare Tokyo residents for a Mount Fuji volcanic eruption

Mount Fuji hasn’t erupted since 1707. But for Volcanic Disaster Preparedness Day, Japanes

29. 8. 2025 14:40:03 | Fast company - tech