Nvidia Unveils New AI Model to Rival ChatGPT

Nvidia has released a new open-source AI model that competes with the best state-of-the-art AI models from the likes of OpenAI, Google, and Anthropic.

The NVLM 1.0 family consists of large multimodal language models, with the largest at 72 billion parameters, exhibiting exceptional performance across vision and language tasks.

Breaking from the trend of keeping advanced AI systems proprietary, Nvidia has released model weights similar to what Meta has done. In addition, Nvidia also released the training code, which will benefit AI researchers and developers.

Open-source AI

It is unclear what the release of such a powerful model will mean for the competitive landscape of AI development, but it certainly opens up new possibilities – potentially pressuring other technology leaders to open their research and accelerate AI progress across the board.

Even if this doesn’t happen, Nvidia’s move could level the playing field for smaller teams and researchers to innovate. This could democratize access to cutting-edge AI technology, fostering a more collaborative and inclusive AI ecosystem.

The AI community has reacted positively to the release, according to VentureBeat.

Cutting-edge performance

Based on the benchmarks shared, Nvidia’s AI model offers better performance and accuracy in vision-language tasks than OpenAI's GPT-4o. It can also interpret data presented in charts and tables, interpret memes, analyze images, and solve complex mathematical equations step-by-step.

Specifically, NVLM accurately performs localization to effectively answer location-sensitive questions such as “What is the difference between the left, middle, and right objects in the image?” and is capable of performing mathematical reasoning and coding based on visual information, such as tables and handwritten pseudocode.

“Our NVLM-D-1.0-72B demonstrates versatile capabilities in various multimodal tasks by jointly utilizing OCR, reasoning, localization, common sense, world knowledge, and coding ability,” the researchers wrote.

Finally, it is worth noting that despite its open-source status, Nvidia has restricted its use to research purposes under its licensing terms. Users are not allowed to use the model for commercial solutions or modify it for resale.

You can learn more about NVLM here.

Image credit:BING-JHEN HONG