Compare byte-level predictions of RWKV7 and Qwen3 models
Interact with a multimodal AI assistant using text and images