DeepSeek dethroned

Alibaba's AI model, called Qwen3, passed R1 in LiveBench tests that assess open-source model capabilities including programming, mathematics and data analysis.

Qwen3, a new technology launched by Alibaba. Photo: SCMP .

According to the latest benchmark tests of the AI ​​world, Alibaba's newly released artificial intelligence model Qwen3 has surpassed DeepSeek's R1 to become the world's highest-ranked open-source model.

Specifically, data from LiveBench, an independent platform that benchmarks large language models (LLMs), the technology that underpins generative AI services like ChatGPT, shows that Qwen3 outperformed R1 in the tests.

The review of open source AI models capabilities includes programming, mathematics, data analysis, and language instruction.

The AI ​​model series called Qwen3 was released by Alibaba on April 28. The company claims that this chatbot can be comparable to, and even surpass, the best models currently available from OpenAI or Google in some cases.

With a size of up to 235 billion parameters, Qwen3 is on par with DeepSeek-V2 and OpenAI GPT-4, which have about 236 billion and 175 billion parameters, respectively. Users will soon be able to download it under an open license on the AI ​​development platform Hugging Face and Github once the series of models is released.

The company says the Qwen3 collection includes hybrid models, meaning they can flex their time to infer to solve complex problems or quickly respond to simple requests. In this case, inference allows the model to self-check the accuracy of information, but incurs high latency.

This design makes it easy for users to allocate the appropriate budget for each specific task. In addition, this model also learns from many competitors around the world.

Using a “mixture of experts” (MoE) architecture similar to DeepSeek, Qwen3 can maximize computational efficiency at a fraction of the training cost. This is an approach that breaks down a task into separate parts and only recommends enough deep data to perform.

According to the development team, Qwen3 supports up to 119 languages ​​and is trained on a dataset of nearly 36,000 billion tokens, equivalent to 27,000 billion words. Training data is taken from many sources such as textbooks, question-answer sets, programming code, or self-generated AI,...

Despite topping the open-source rankings, extensive testing by LiveBench shows that Qwen3 lags behind the world’s leading closed-source AI models, notably OpenAI’s o3, Google’s Gemini Pro 2.5, and Anthropic’s Claude 3.7.

Currently, OpenAI's top-of-the-line model backed by Microsoft, o3-mini high, is at the top of the overall AI model rankings in the world.

Source: https://znews.vn/deepseek-bi-soan-ngoi-post1551500.html