Open AI, the makers of one of the most potent AI tools GPT-4o recently announced an upgrade to the version. The company released a fine-tuning feature for custom datasets, which will help developers extract the complete processing abilities of the LLM, while also pushing its limitations through customization of datasets. This can help the developers improve efficiency and obtain more accurate results.
GPT-4o, launched in May 2024 is available to users on a subscription basis. With the launch of its new upgrade, Open AI said it would provide free training programs to companies and developers to improve and add more customization to the model. Starting from the second half of September, the makers assured the organizations using GPT-4o and GPT-4o mini will get one million and two million training tokens per day respectively. For other users, it can cost $25 per million tokens, which is roughly around INR 2000.
The makers emphasized directly working with companies through their GPT-4o model, which eventually lowers the need for businesses to outsource or obtain external services. The organizations should be subscribed to the Open AI, where they will continue by uploading their data to the Open AI servers. The training period will last for a maximum of 2 hours, after which the model can be used.
It should be noted that this model is specifically targeted at text data and does not include images or other forms of content. However, the makers displayed optimism towards bringing new upgrades in the future, which might also include images and other content forms, thus making it a multi-purpose model.
Open AI has already cemented a robust position in the AI industry. With its new model GPT-4o, the AI market seems heading to nowhere but advancement. Moreover, the multimodality of the model allows the user to handle the tasks on a single platform.
The progressive and inclusive approach with its new LLM model will allow developers and organizations to harness the potential of AI in their tasks, thus enhancing the accuracy and quality of the work.