ChatGPT has been "overtaken" and is no longer the smartest AI chatbot today

9V2C...mFQ8
3 Apr 2024
37

For a long time, ChatGPT is still considered the smartest AI-integrated chatbot thanks to its large database and fast user response speed. However, recently this position of ChatGPT has been "overthrown

ChatGPT - chatbot (automatic chat software) integrated with artificial intelligence (AI) - has suddenly become a global phenomenon, when this chatbot demonstrates incredible intelligence through social content. dialogue with users

Users can ask questions and communicate with ChatGPT in text, this tool will provide answers and follow user requests, such as writing paragraphs, writing programming code, composing emails... with very natural language, like a real human being. Notably, ChatGPT supports many different languages, including Vietnamese.

Since ChatGPT created a "fever" globally, many large technology companies have quickly started building AI-integrated chatbot tools to compete with ChatGPT, opening up a race for development. AI, such as Google's Gemini, Alibaba's Qwen, Microsoft's Copilot or Meta's Llama

However, ChatGPT still firmly holds the "throne" in the race and is still considered by both technology and users as the smartest AI chatbot today.

However, recently, ChatGPT is no longer the smartest AI chatbot in the world, according to the just-published rankings of LMSYS, an organization specializing in evaluating and ranking the capabilities of major language models. foundation for developing AI chatbot tools.

According to LMSYS's "Chatbot Arena" rankings, the large language model Claude 3 Opus, developed by San Francisco-based startup Anthropic, has surpassed OpenAI's GPT-4-1106-preview to to become the world's smartest large language model

Claude 3 Opus is the language model used to develop the Claude AI chatbot, while GPT-4 is being used as the foundation for OpenAI's ChatGPT professional version chatbot. This is the first time OpenAI's language model has been knocked off the top spot since LMSYS launched the "Chatbot Arena" rankings a year ago. This ranking is constantly updated and the ranking of language models is always mixed, but OpenAI's GPT has never left the first position, until now.
The GPT-4 language model with a lower version (GPT-4-0125-preview) ranked 3rd in the LMSYS rankings, while the Bard language model (used to develop the Gemini chatbot Professional version) of Google ranked 4th in terms of intelligence. Notably, the rating scores for the 3 major language models leading in the rankings differ very little, which shows that the intelligence level of chatbots built on these 3 language models is equivalent.

Among the 10 major language models ranked first in intelligence, Anthropic contributes 3 names, including Claude 3 Opus, Claude 3 Sonnet (4th place, used for Claude AI free version), and Claude 3 Haiku (first Claude 3 version, ranked 7th).

Meanwhile, OpenAI has 4 products in the top 10 smartest language models, with 2 test versions of GPT-4 (preview version, ranked 2nd and 3rd respectively), GPT -4-0314 (ranked 6th) and GPT-4-0613 (ranked 8th).

The top 8 positions in the top 10 smartest major language models all belong to American companies. The Mistral-Large-2402 language model of French technology company Mistral and Qwen1.5-72B-Chat of Chinese technology company Alibaba appeared in 9th and 10th positions, respectively.

Experts predict that when OpenAI launches a completely new GPT-5 language model, with many improvements compared to the current GPT-4, ChatGPT will soon return to the leading position in the AI-integrated chatbot race.
LMSYS (Large Model Systems) Organization is a research organization founded by AI experts at the University of California Berkeley, University of California San Diego and Carnegie Mellon University to research AI systems and evaluate them. major language models.

"Chatbot Arena" is a leaderboard created by LMSYS to evaluate and rank the most popular and widely used major language models today. In addition to reviews from experts, "Chatbot Arena" also records reviews from the user community when using chatbots in practice.


BULB: The Future of Social Media in Web3

Learn more

Enjoy this blog? Subscribe to langthang

1 Comment