Nvidia’s new AI model released to compete with GPT-4

Nvidia A family of open source and powerful artificial intelligence models NVLM 1.0 It has released itself to compete with models made by OpenAI and Google. The main member of this family of large multimodal language models named NVLM-D-72B has 72 billion parameters It provides exceptional performance in visual and language tasks, and its text capabilities have also increased.

Nvidia researchers explain in their article:

“We introduce NVLM 1.0, a family of large multimodal language models that achieve state-of-the-art results on vision tasks and compete with leading models (such as GPT-4o) and other available models.”

Features of Nvidia’s new artificial intelligence model

According to published reports, the NVLM-D-72B model has shown excellent adaptability in processing complex visual and textual inputs. In the Nvidia article, examples are presented that show the model’s ability to interpret memes, analyze images, and solve math problems step by step.

In addition, the performance of the model has also been improved in the field of textual tasks. In fact, NVLM-D-72B’s accuracy has improved by an average of 4.3 points in key benchmarks.

With the public release of this model and the promise of releasing its training code, Nvidia apparently intends to break away from the trend of keeping advanced artificial intelligence systems closed. A decision that gives researchers and developers access to advanced technologies and accelerates research and development in the field of artificial intelligence.

In addition, the NVLM project includes innovative architectural designs, including a hybrid approach that combines different multimodal processing techniques. Such progress can shape the direction of future research.

The artificial intelligence community has generally shown a positive reaction to Nvidia’s new artificial intelligence model. One of the artificial intelligence researchers in the social network X says:

“Nvidia just released the 72B, which is on par with the Llama 405B 3.1 in math, coding, and vision capabilities.”

However, the public release of NVLM 1.0 has its own risks. As more powerful AI models become available, concerns about their misuse are likely to increase.