libaba, the Chinese tech company released this Wednesday a new and improved version of their artificial intelligence model called Qwen 2.5, and claims that it
has more powerful capabilities than most of the AI models on the market.
Qwen2.5-VL can support various texts and image analysis and also can understand videos, and enumerate objects from images. Maybe the most impressive feature of this new Alibaba AI model is that is capable of fully controlling a PC, a feature that is similar to the Operator launched recently by OpenAI.
According to the Qwen team’s benchmarking, the new Alibaba AI model surpasses the rival models in almost every measure. Placed next to the OpenAI’s GPT-4o, Claude 3.5 Sonnet from Anthropic, and Google’s Gemini, Alibaba’s Qwen2.5-VL recorded some remarkable scores.
It should be mentioned that the less advanced models released by Alibaba, Qwen2.5-VL-3B and Qwen2.5-VL-7B can be accessed and explored under a permissive license. On the other hand, their artificial intelligence flagship, Qwen2.5-VL-72B is currently under a custom license, and companies need to require Alibaba’s permission in order to have and use their most powerful AI model.
This new release took place on an unusual day, the first day of the Lunar New Year when most Chinese people celebrate with their families. Maybe, the pressure from other AI models such as DeepSeek has pushed Alibaba to release their new model sooner than was initially planned.
Subscribe to our newsletter
The AI startup has experienced a significant rise in popularity due to its impressive artificial intelligence capabilities on the R1 and V3 models and has managed to raise concerns among industry competitors which are somehow forced now to release more advanced artificial intelligence chatbots.
"Qwen 2.5-Max outperforms ... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B,", the cloud unit from Alibaba stated in their announcement on WeChat account.
DeepSeek AI startup launched the V3 model this month on January 10, while the R1 model was released 10 days later, on January 20. These two models left all the investors impressed by their features and capabilities and determined some of them to consider a potential investment in the future.
Currently, DeepSeek managed to surpass OpenAI’ChatGPT in the Apple App Store, becoming the top app based on the productivity criteria.
While there are not yet tests made between the new Alibaba AI model and DeepSeek’s V3 and R1 chatbots, based on the scores we can say that these two are currently the most powerful models that use artificial intelligence technology.
Stay tuned for more updates!
By
George King
•
January 29, 2025 11:40 AM
.webp)
.webp)
.webp)
