The field of geometry reasoning and mechanism design is witnessing significant developments, with a focus on improving the performance of multimodal large language models (MLLMs) in solving complex geometric problems. Researchers are exploring new evaluation strategies, such as the use of auxiliary lines, to assess the long-step reasoning abilities of MLLMs. Additionally, there is a growing interest in developing topology-aware reasoning frameworks that can efficiently learn and generate diverse topological structures. The application of sequence learning and Transformer-based models is also being investigated for mechanism design automation, with promising results in generating novel and valid designs. Noteworthy papers in this area include: STELAR-VISION, which introduces a training framework for topology-aware reasoning and achieves state-of-the-art results on several benchmarks. MechaFormer, which presents a Transformer-based model for mechanism design automation and demonstrates significant improvements in path-matching accuracy and design diversity. Bridging Formal Language with Chain-of-Thought Reasoning to Geometry Problem Solving, which explores a new approach that integrates Chain-of-Thought with formal language and achieves up to 15% accuracy improvements on standard GPS benchmarks. CAD-RL, which proposes a multimodal Chain-of-Thought guided reinforcement learning framework for CAD modeling code generation and achieves significant improvements in reasoning quality, output precision, and code executability.