Anthropic Claude 3 Big model test results were released: A new round of GPT revolution started

On March 5, OpenAI’s strongest rival Anthropic released its Claude 3 series models, with three high-performance models, arranged in ascending order of capabilities: Claude 3 Haiku, Claude 3 Sonnet, Claude 3 Opus,

And announced that it has surpassed ChatGPT-4 in its AI logic benchmarks.

Anthropic claims that Claude 3 is also a multimodal large model with strong “visual capabilities”, so users can upload photos, charts, documents and other types of data to analyze and ask questions.

Claude 3 Far outperformed OpenAI’s GPT-4 and Google (GOOG) Gemini Ultra in industry benchmarks, such as knowledge level, deductive reasoning and basic mathematics.

According to the test data released by Anthropic, the capability of its highest capability model Claude 3 Opus has completely surpassed the GPT-4. Meanwhile, Anthropic also revealed that it plans to release frequent updates to its Claude 3 model series in the coming months.

With the popularity of ChatGPT, OpenAI, popularity exploded. In fact, Anthropic is a company founded by former OpenAI employees and is considered OpenAI’s biggest competitor. Over the past year, it has received a $4 billion investment from Amazon (AMZN).

Last year, the AI model competition exploded around the world. Microsoft (MSFT) is far ahead in the competition with Amazon and Google in the field of artificial intelligence by binding Open AI GPT model. Then, with the emergence of Claude 3 Opus, which comprehensively crushed GPT-4, Amazon will continue to conduct in-depth research and development in this AI game.

With the release of the Anthropic Claude 3 series models, Amazon said that Anthropic’s latest Claude 3 series models will be available on Amazon Bedrock. With Amazon Bedrock, customers will be able to easily use the latest model of Anthropic, and in the past year, after enduring the pressure of Microsoft’s cloud service growth, Amazon seems to pull back!

It has to be said that with the release of the strongest AI large model Claude 3, it has once again attracted the attention of the AI industry. With the leapfrog development of multimodal large model, Multimodal large language model (MLLM) has shown unprecedented ability in processing multimodal content such as text, code, images and video, and will become the mainstream wave of technology.

The large language model (LLM) represented by Llama 2 and Mixtral, and the multi-modal large language model represented by GPT-4, Gemini and LLaVA have developed by leaps and bounds. In the New Year, the field of artificial intelligence has multi-modal AI function —— that is, to add photos, videos and other input options —— whether users upload or through the AI model creation, have quickly become a hot spot in the industry.

At present, the global competition in the field of large models is still fierce, which will continue to improve the overall ability level of large models. According to the data, WiMi Hologram Cloud (NASDAQ: WIMI), a leading AI vision enterprise, with the breakthrough of the big model, based on artificial intelligence technology capabilities quickly penetrated into live broadcasting, online tourism, education industry, online advertising and other fields. For now, WiMi has increased significantly in its technical, large-model, multimodal and data understanding aspects.

In fact, ChatGPT mainly focuses on text understanding, but today’s big models focus on multimodal understanding. Multimodal models can process a variety of media data and integrate them into a unified semantic space. For example, GPT-4 can read pictures, answer questions, reason with data, analyze charts, etc. The multi-modal capability of GPT-4 will surely generate a wider range of downstream applications.

And WiMi on the footsteps of chasing big model, always adhere to their own development strategy, in terms of AI big model foundation ability, around the language, knowledge, understanding, algorithm, multimodal as the core, including content evaluation and summary, multilingual translation, using its advantages to expand AI skills and creativity, through technical means to process video, pictures, audio, etc., to automated production ideas, meet the demand of the market diversification.

In addition, large models require large computing power, and the iterative upgrade of large model algorithms will provide a strong impetus for the growth of the global and AI computing power markets. In the face of the complex competitive environment, WiMi will accelerate the research and development work of the AI chip industry chain, and its related AI chip and scenario application may be deeply benefited. In the future, under the multi-application end, WiMi multi-modal large model capability will continue to upgrade, and the development prospect around the layout of AI industry is broad.

The AI trend is unchanged, and it’s not surprising that Anthropic claims the ability to surpass GPT-4. However, the launch of the Claude 3 does intensify the competition for the big AI models, and the industry is now looking to the launch of the GPT-5. While Anthropic executives claim that Claude 3 is the most powerful big model yet, once OpenAI GPT-5 is released, there will undoubtedly be a new AI battle and come close to the AGI era.