Vibeleaderboard
← Benchmark guide

Model results

BullshitBench

Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.

Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.

Evaluation groups

Open a group to see its exact protocol and scores. Different configurations remain separate.

Model rankings

Open a group to compare scores within its reported protocol. Positions do not establish statistical significance.

  • BullshitBenchv2

    BullshitBench (reviewed publisher results)

    2 models with scores

    Measurement date not reported

    Scope · Clear pushback rate; 2/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 2/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • BullshitBenchv2

    BullshitBench (reviewed publisher results)

    2 models with scores

    Measurement date not reported

    Scope · Clear pushback rate; 8/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 8/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • BullshitBenchv2

    BullshitBench (reviewed publisher results)

    2 models with scores

    Measurement date not reported

    Scope · Clear pushback rate; 10/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 10/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • BullshitBenchv2

    BullshitBench (reviewed publisher results)

    2 models with scores

    Measurement date not reported

    Scope · Clear pushback rate; 83/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 83/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

Other reported results

These reports have no model ranking. Open a group for its evidence and the reason positions are withheld.

  • Claude Fable 5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 45/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 45/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=low; result date not supplied; source captured 2026-09-10

  • Claude Fable 5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 56/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 56/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • Claude Haiku 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 71/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 71/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Haiku 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 77/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 77/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Opus 4v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 34/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 34/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.1v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 42/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 42/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.1v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 43/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 43/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 79/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 79/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 90/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 90/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.6v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 87/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 87/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.7v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 74/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 74/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=max; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.8v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 94/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 94/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • Claude Opus 4.8v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 95/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 95/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Opus 5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 70/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 70/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • Claude Opus 5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 73/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 73/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=low; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 4v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 29/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 29/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 4v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 30/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 30/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 74/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 74/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 79/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 79/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 4.6v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 89/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 89/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 4.6v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 91/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 91/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 78/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 78/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=max; result date not supplied; source captured 2026-09-10

  • Claude Sonnet 5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 80/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 80/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=low; result date not supplied; source captured 2026-09-10

  • ERNIE 4.5 VL 424B A47Bv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 3/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 3/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • ERNIE 4.5 VL 424B A47Bv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 11/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 11/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • Gemini 2.5 Flashv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 19/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 19/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • Gemini 2.5 Prov2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 20/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 20/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • Gemma 3 27Bv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 3/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 3/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • GLM 4.5v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 6/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 6/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • GPT-4o (2024-08-06)v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 12/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 12/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • GPT-5 Codexv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 39/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 39/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=default; result date not supplied; source captured 2026-09-10

  • gpt-oss-120bv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 5/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 5/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • gpt-oss-120bv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 11/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 11/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=low; result date not supplied; source captured 2026-09-10

  • MiMo-V2-Flashv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 13/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 13/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • MiMo-V2-Flashv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 16/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 16/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Mistral Small 4v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 4/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 4/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Mistral Small 4v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 6/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 6/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=high; result date not supplied; source captured 2026-09-10

  • R1v2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 7/100; all three judges; rounded consensus=2; errors and refusals included in d…

    Clear pushback rate; 7/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=none; result date not supplied; source captured 2026-09-10

  • Step 3.5 Flashv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 13/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 13/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=minimal; result date not supplied; source captured 2026-09-10

  • Step 3.5 Flashv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 22/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 22/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • Trinity Large Thinkingv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 13/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 13/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=xhigh; result date not supplied; source captured 2026-09-10

  • Trinity Large Thinkingv2

    BullshitBench · BullshitBench (reviewed publisher results)

    Single-model result · Positions withheld

    Measurement date not reported

    Scope · Clear pushback rate; 17/100; all three judges; rounded consensus=2; errors and refusals included in …

    Clear pushback rate; 17/100; all three judges; rounded consensus=2; errors and refusals included in denominator; reasoning=minimal; result date not supplied; source captured 2026-09-10

BullshitBench model results | VibeLeaderboard