Performance on General Reasoning Benchmark

[SxJyJay]

Thanks for your great work.

I do not find the performances of OpenCUA series on general reasoning benchmarks like MMMU and MathVision. What are their performances compared to the Qwen2.5-VL-7B-Instruct.

[XingWang1234]

I think this model is specially trained for computer use task, similar to OpenAICUA model. The aim is not a general model.