openai has problem meeting demand for what gpt-4 in its current form is capable of. Adding features like pics or videos would create even more strain on the system so even if this model has additional capabilities and it probably does, they simply need some time to make it work for such high traffic