Internal court filings reveal that Meta's executives and researchers were fixated on surpassing OpenAI's GPT-4 while developing their own Llama 3 model. Ahmad Al-Dahle, Meta’s VP of Generative AI, emphasized the need to outdo competitors like Anthropic and OpenAI in a series of messages. The filings suggest a dismissive attitude towards rivals, such as the French startup Mistral. Amidst the pressure to release advanced AI models, concerns have emerged regarding the legality of the data used for training. CEO Mark Zuckerberg has stated that Meta aims to close the performance gap with leading AI models by 2025.