On prompt tokens, Qwen3-VL’s selected routing weights are substantially flatter than the text model’s: its top three hold less of the selected-eight mass, and the selected eight hold less of the full 128-way softmax mass.

On prompt tokens, Qwen3-VL’s selected routing weights are substantially flatter than the text model’s: its top three hold less of the selected-eight mass, and the selected eight hold less of the full 128-way softmax mass. · CiteArk

Measurement results

0/6 assessed

0 supported0 disputed or conflicting6 not assessed

The complete result group stays with its claim. Assessed rows come first; their judgments do not replace the whole-claim conclusion. Expand a row for conditions, sources, and evidence history.

Each row is a measurement with its reported value, latest available observation, and cumulative evidence state.
Metric and conditions	Reported	Latest observation	Evidence judgment
printedValue: 49.8	49.8%	—	Not assessed
printedValue: 59.8	59.8%	—	Not assessed
printedValue: 50.2	50.2%	—	Not assessed
printedValue: 40.2	40.2%	—	Not assessed
printedValue: 15.5	15.5%	—	Not assessed
printedValue: 44.3	44.3%	—	Not assessed

On prompt tokens, Qwen3-VL’s selected routing weights are substantially flatter than the text model’s: its top three hold less of the selected-eight mass, and the selected eight hold less of the full 128-way softmax mass.

Reported and observed measurements

Assessments (0)

Runs (0)

Measurement results