After Deepsek, another Chinese company’s AI model did wonders, beaten Openai, like this

After Deepsek, another Chinese company’s AI model has rescued the sweat of American companies. Actually, the AI ​​model of China’s Startup Moonshot AI Kimi K1.5 has surpassed Openai’s GPT-4o and Anthropic’s Claude 3.5 Sonnet model on several benchmarks. It has also managed to compete with Openai-01 in understanding maths, coding and text and visual etc. It has also cost a very low cost like Deepsek-R1.

Kimi K1.5 is not just AI model

Kimi K1.5 is not just an AI model. It is considered to be a big bounce in reinforcement learning and multimodal region. It is capable of solving any complex problem by mixing visual data, text and code. It has defeated models of American companies in many benchmarks. It has been trained in reinforcement learning methods. Where the rest of the AI ​​models depend on static dataset, it learns from exploration and refining. This increases the ability to solve complex issues.

How does this work?

According to media reports, this model works using Chan of Thoughts Approach. This means that it breaks a difficult problem in small steps. Then it tries to solve them by using regional on them. Since it can process text and image, it can be used to solve text-image analysis and the problem that requires visual input.

Deepsek has created Tehelka

Chinese startup Deepsek has created panic in the world with its AI models. Within a week, it has left behind Openai’s chatgpt on the app store in case of free apps. It has been very low to make it, which has surprised the Tech world.

Also read-

Shocked the world by bringing cheap AI models, know the full story of Chinese Startup Deepsek, which causes earthquake in the tech world