Skip to main content
Radar

Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.

Start free brief

Benchmark profile

MASK-Rectified dishonesty rate (MASK-Rectified dishonesty)

A rectified dishonesty rate derived from the public MASK honesty benchmark family.

Data verified 26 confirmed releases in the last 30 daysStart free brief

Benchmark score on MASK-Rectified dishonesty — August 12, 2026

BenchLM mirrors the published score view for MASK-Rectified dishonesty. Grok 4.5 leads the public snapshot at 0.67% , followed by Grok 4.6 (3.80%). We do not use these results to rank models overall.

2 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated August 12, 2026

Benchmark score table (2 models)

Score
1
Grok 4.5xAI · Closed
0.67%
2
Grok 4.6xAI · Closed
3.80%

About MASK-Rectified dishonesty

Year

2025

Tasks

Pressure-to-lie scenarios

Format

Dishonesty rate

Difficulty

Model honesty

MASK publishes a paper, dataset, and code. BenchLM keeps xAI's rectified implementation separate and display-only; lower dishonesty is better.

BenchLM freshness & provenance

Version

MASK-Rectified dishonesty 2025

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does MASK-Rectified dishonesty measure?

A rectified dishonesty rate derived from the public MASK honesty benchmark family.

Which model scores highest on MASK-Rectified dishonesty?

Grok 4.5 by xAI currently leads with a score of 0.67% on MASK-Rectified dishonesty.

How many models are evaluated on MASK-Rectified dishonesty?

2 AI models have been evaluated on MASK-Rectified dishonesty on BenchLM.

Compare Top Models on MASK-Rectified dishonesty

Last updated: August 12, 2026 · BenchLM version MASK-Rectified dishonesty 2025

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.