REASONING VLM FROM KIMI

Summary from the TLDR AI newsletter

One of the first open reasoning models for visual question answering. This small model is extremely good at grounded question answering for images, image captioning, and even some math based on image input.