OpenAI is actively working on the development of Gobi, an advanced large language model with multimodal capabilities. The goal is to stay ahead of Google's Gemini. In our previous news article, it was discussed that Gemini, Google's upcoming multimodal model, is expected to be released in December 2023. OpenAI plans to add similar multimodal functionality to GPT-4 to compete effectively.
The rise of multimodal language models
Multi-modal language models are currently the talk of the town. ChatGPT is a notable example of their performance across various domains. These models use large language models as their 'core' for diverse multi-modal tasks. Think of generating stories based on images, answering questions about visual information, and performing mathematical reasoning without OCR.
OpenAI already demonstrated these capabilities in March with GPT-4, albeit with limited access, available only to one company called "Be My Eyes," which develops mobile applications for the visually impaired. Now, OpenAI is preparing to roll out GPT-Vision on a larger scale, six months after the initial demonstration.
The delay in the rollout is primarily due to concerns about potential misuse by malicious actors, which OpenAI's engineers are currently addressing.
Google faces similar challenges and made commitments in July to ensure the responsible development of all their products and to prevent misuse of Gemini. Despite these challenges, Google has an advantage due to their own data on text, images, videos, and audio, including data from platforms like search queries and YouTube. Early users of Gemini have already reported fewer incorrect answers compared to existing models.
Plans for GPT-5 and Future Developments
Sam Altman, CEO of OpenAI, recently stated that GPT-5 has not yet been released. However, they plan to enhance GPT-4 with various improvements, one of which may potentially become the new upgraded model. It is still too early to say whether Gobi will ultimately become GPT-5, as the training process does not seem to have begun yet.
This competition is reminiscent of the rivalry between iPhone and Android in the world of artificial intelligence. People are looking forward to the launch of Gemini, which will reveal the competition between Google and OpenAI and the potential impact this may have on businesses.