Skip to content
Trending storyDeveloping

Google and Artificial Analysis launch Multimodal Knowledge Agent benchmark

1 article1 sourcesince Nov 10Last article 1h ago ·

Overview

AISummary of 1 article

Google and Artificial Analysis launch MKA, a benchmark that tests how AI agents reason across visually complex files on real-world knowledge work tasks.

Artificial Analysis runs a 200-task evaluation set that stays private so the tasks cannot leak into training data or be found through search. Google also releases a separate 10-task public set on Hugging Face so anyone can inspect the format and test their own agents.

Google says Gemini 4 Carbon ranks second in quality among the tested models at 6x lower average cost than the top scorer.

Written by AI from the articles below · updated Nov 10, 12:04 PM ET

Check the sources:

Latest developmentNov 10, 11:00 AM ET
Google and Artificial Analysis launch the Multimodal Knowledge Agent benchmark

Article timeline

The articles in this story. Times are ET.

Nov 10
  1. Google · Gemini appOfficial
    Google and Artificial Analysis launch the Multimodal Knowledge Agent benchmark

    AIGoogle and Artificial Analysis launch MKA, a benchmark that evaluates how agents reason across visually complex files on real-world knowledge work tasks. The 200-task evaluation set is kept private, and a separate 10-task public set is released on Hugging Face. Google says Gemini 4 Carbon ranks #2 in quality among the tested models at 6x lower average cost than the top scorer.

Heat Trend

Not enough continuous observations to show a trend yet.