Four-Factor Integrity Breakdown & Primary-Source Evidence
Access Model: Strictly Private
"Evaluation set: We run a 200-task private set across 10 industry verticals. Each task has a prompt, a set of input files (PDFs, Office documents, spreadsheets and images), and a rubric written by domain experts."
Originality
Weight 25%
10/10
Reuse Profile: All Original (0% reuse)
"Artificial Analysis' implementation of Google DeepMind's Multimodal Knowledge Agent (MKA) benchmark. It tests whether agents can read visually dense project files, such as architectural drawings, engineering schematics and clinical scans, and produce the deliverables a professional would."
Freshness
Weight 20%
10/10
Initial Release: 2026-11Latest Refresh: 2026-11Rolling Updates: No
"November 10, 2026 Multimodal Knowledge Agent (MKA) benchmark: testing knowledge work capabilities with complex visual files"
Grading
Weight 15%
3.3/10
Integrity: Self-Reported / Live NetworkSandbox: live internet
"The agent works in a Linux sandbox that holds the input files and has internet access. It has a code execution tool, web search and page-fetch tools, and an image-viewing tool that we enable only for models that support image inputs."
Frontier Model Standings on MKA (Multimodal Knowledge Agent)
Top 15 active models tracked
Benchmark maintainers can embed this live SVG badge in GitHub READMEs or documentation:
[](https://benchaudit.org/benchmarks/mka/)