Part 4/11:
Despite its innovative reasoning approach, Strawberry has notable limitations. Currently, it will only process text-based queries—a step back from GPT-4’s multimodal capabilities, which include image and video processing. Some early testers have also noted that Strawberry’s response times for simpler questions might feel unnecessarily slow, which could frustrate users expecting swift replies.
Additionally, while Strawberry is designed to remember past conversations for more personalized interactions, it has struggled with maintaining consistency over longer dialogues. This remains an area under evaluation, as OpenAI seeks to optimize the model’s contextual understanding without sacrificing user experience.