Multimodal LLM benchmarks: MMIE, HumanEval-V, MixEval-X. AI video generation: MovieGen, PyramidFlow. Multimodal LLMs: Emu3, MIO, MM-1.5. DepthPro for depth perception.