Claude 3 was released, surpassing GPT 4 in terms of performance丨Nvidia became the third largest comp

Mondo Technology Updated on 2024-03-07

AI Singularity丨AI News丨AI Tool Recommendation丨AI Tool Tutorial.

Provide 1000+ AI efficiency tools丨Welcome to pay attention.

AI Singularity.com March 5, 2024 ** Dao丨 AI Information Morning Post

On March 4, local time, Anthropic, an artificial intelligence startup that is considered to be one of OpenAI's strongest competitors, officially and quietly released a new generation of Claude large models - Claude 3 series multimodal large models on its official website.

It is reported that Claude 3 has three versions: Haiku, Sonnet and Opus. Judging from the official evaluation report, according to its evaluation report, OPUS surpasses OpenAI's GPT-4 in 10 dimensions such as graduate level expert reasoning, basic mathematics, undergraduate level expert knowledge, and **.

In addition, the Haiku model is more efficient, and can read a 10,000 tokens in 3 seconds. Sonnet 2 than the previous Claude 2Version 1 is more intelligent and suitable for tasks such as knowledge retrieval.

All three models currently support 200,000 context windows. Anthropic says they also support 1 million contexts, but are only available to specific users due to the huge amount of AI computing power required.

Previously, users can use the Claude 3 Sonnet model for free; If you want to use the most powerful version of Opus, you need to open a membership; The Haiku model is coming soon. At the same time, the Claude API is widely used in 159 countries.

On February 24, local time, OpenAI announced that it would close the beta version of the ChatGPT plug-in from March 19, the purpose of which is to guide developers and users to focus on the GPTS service launched at the end of last year.

As of March 19, users will not be able to create new conversations using the plugin, but existing conversations will be retained until April 9 at the latest.

According to official information, during the plugin testing period, there were already more than 1000 plugins available in the plugin store. There are already thousands of GPTs in the GPT store, covering areas such as writing, productivity, programming, education, and more. OpenAI claims that GPTS already has full parity with plugins, which is one of the main reasons for the closure.

On March 1, Meizu released its new smartphone product, the Meizu 21 Pro, which Meizu called its first "open AI terminal" and announced that it would be open to all large model teams.

It is understood that the new product supports a number of AI end-side capabilities: such as AI smart keys, AI auxiliary input, AICY voice assistant, AI gallery and other AI functions.

According to official information, AICY voice assistant has the ability of copywriting, image creation, etc., the gallery application supports natural language search, AI smart keys can support AI screen recognition and other functions, and AI input assistance can generate reply suggestions for long article creation.

On March 1, local time, Microsoft released an update to the new version of Windows 11, focusing on upgrading and improving Copilot, bringing AI functions to more system-level integration.

Copilot gets more features to control the user's computer, such as turning on the PC's power-saving mode, enabling accessibility features such as Narrator or Magnifier, displaying available WiFi networks, cleaning up stations, and more. The Windows 11 update will be available from today, but Copilot's new features will be available from late March.

On March 4, local time, the market value of artificial intelligence chip giant Nvidia rose sharply at the opening, officially surpassing Saudi Aramco and becoming the third highest listed company in the world by market value after Microsoft and Apple. According to the data, as of last Friday, the market capitalization of Nvidia reached 206 trillion US dollars, crossing the 2 trillion US dollar mark for the first time, while Saudi Aramco's market capitalization fell to 203 trillion dollars.

Nvidia's spurt of earnings has once again confirmed Wall Street's optimism about the potential of artificial intelligence to reshape the tech industry, which has sent Nvidia's stock price soaring. After rising more than 239% last year, Nvidia has maintained a strong rally this year, with a year-to-date increase of more than 66%, and its market value has increased by about $834 billion.

In terms of core business, according to the fourth-quarter financial report, Nvidia's data center business revenue increased by 409% year-on-year to $18.4 billion, mainly driven by strong demand for large models, recommendation engines and generative AI.

Google is optimistic about the prospect of applying large AI models to smartphones. A few days ago, Brian Lakovski, vice president of product management in Google's pixel division, told ** that Google is ready to embed more advanced Gemini large models into the next generation of smartphones, including partners, such as Samsung smartphones.

[**丨ai singularity network丨The whole network account has the same name丨Welcome to follow].

AI Singularity丨Provide 1000+ AI efficiency tools丨Welcome to pay attention.

Related Pages