RLTasksby ulam.ai
Models / Qwen

Qwen / MODEL EVIDENCE

qwen/qwen3.5-397b-a17b@reasoning=high

1 benchmarks · 4 observations · 2 sources

Scores at a glance

1 benchmarks · large dot: qwen/qwen3.5-397b-a17b@reasoning=high; small dots: closest models where reported

BullshitBench / SlopBench ↗52.73%
3.6496.36

qwen/qwen3.5-397b-a17b@reasoning=high · BullshitBench / SlopBench · BullshitBench V1 · Clear-answer rate · V1 ↗

3 more reported results

1.3455 / 2 · qwen/qwen3.5-397b-a17b@reasoning=high · BullshitBench / SlopBench · BullshitBench V1 · Mean judgment score · V1 · BullshitBench V1 — published model leaderboard ↗

78% · qwen/qwen3.5-397b-a17b@reasoning=high · BullshitBench / SlopBench · BullshitBench V2 · Clear-answer rate · V2 · BullshitBench V2 — published model leaderboard ↗

1.7033 / 2 · qwen/qwen3.5-397b-a17b@reasoning=high · BullshitBench / SlopBench · BullshitBench V2 · Mean judgment score · V2 · BullshitBench V2 — published model leaderboard ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. BullshitBench V1 — published model leaderboard ↗

    Peter (petergpt) · 2026

  2. BullshitBench V2 — published model leaderboard ↗

    Peter (petergpt) · 2026

Names used in source tables

qwen/qwen3.5-397b-a17b@reasoning=high