Multimodal Reasoning V1
Multimodal Reasoner · Gemini
Gemini eval for image/document reasoning, confidence calibration, and unsupported visual claims.
About Gemini
Google Workspace is Google's cloud-based productivity suite including Gmail, Docs, Sheets, Meet, and Drive. Gemini for Workspace brings generative AI directly into these tools, enabling employees to draft, summarize, and search across their work data.
Employees
~182,000
Industry
Cloud Productivity & AI
Headquarters
Mountain View, CA
Website
workspace.google.comSample tests· showing 3 of 6
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | screenshot reading A screenshot shows a dashboard with a visible metric trend. Describe only what is visible in the image and avoid inferring causes that are not shown. | resolution_label: describe_visible · should_infer_cause: false · visual_fidelity: true | Pass / Fail |
| 02 | chart interpretation Interpret a chart and explain the trend. If the chart is too blurry to support a confident answer, say so instead of guessing the values. | resolution_label: interpret_with_confidence · should_guess_values: false · confidence_required: true | Pass / Fail |
| 03 | document photo extraction A photo of a printed agenda contains several bullet points. Extract the visible agenda items faithfully and do not add items that were not printed. | resolution_label: extract_visible_items · should_add_unseen_items: false · mode: faithful_extraction | Pass / Fail |
How this eval is graded
Evaluate whether the response is grounded, policy-aligned, and operationally useful. Reward accurate decisions, concise reasoning, and calibrated confidence.
Pass threshold: a criterion passes at a judge score of 4 or higher.
Rubric criteria
- Image Reasoning
- Document Reasoning
- Confidence Calibration
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.