Reddit user tests MKA benchmark across six AI workflows
A Reddit submitter tested an Education/History task from Hugging Face against Google and Artificial Analysis's Multimodal Knowledge Agent (MKA) benchmark rubric. The six AI workflows included GPT, Gemini, Claude, and NotebookLM; some produced polished presentations but struggled with numerical accuracy. The submitter published the results, actual answers, and observations in a separate blog.
The hands-on comparison offers a concrete example of how different AI workflows perform on one multimodal knowledge task, while its single-task scope limits the strength of the findings.
Google and Artificial Analysis recently introduced the Multimodal Knowledge Agent (MKA) benchmark, and I thought it would be interesting to try it myself! I downloaded an Education/History task from Hugging Face and tested it across six AI workflows, including GPT, Gemini, Claude, and NotebookLM.
I compared their outputs against the grading rubric, and the results were quite interesting. Some tools created great-looking presentations but struggled with getting the numbers right.
I put together a small blog sharing the results, actual answers, and a few things I noticed.
Here's what I found
#Google#Artificial Analysis#Hugging Face#Multimodal Knowledge Agent (MKA)#GPT#Gemini#Claude#NotebookLM