BLINK

What it measures

14 classic computer-vision perception tasks (relative depth, visual correspondence, forensic detection, multi-view reasoning, jigsaw) solvable by humans "in a blink" (<1s) that stump VLMs.

Reading the measurement

Study what information the model receives and which inference the answer requires. A question may mix recall, visual perception, deduction and search. Use task breakdowns to find the mechanism behind a headline gain.

Published evaluation

Test accuracy; BLINK · 1,907 test items. The table retains the source’s reported model and evaluation configuration.

Where this fits in an evaluation suite

Ask what the task actually requires: recall, deduction, search, or a mixture. Separate closed-book reasoning from retrieval and tool-assisted problem solving.

Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.

A reproducible comparison

Test accuracy; BLINK · 1,907 test items. The table retains the source’s reported model and evaluation configuration.

Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.