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

HackerBench v0.2 harmful and dual-use compliance (HackerBench harmful compliance)

The rate at which a model complies with harmful or dual-use cyber requests in xAI's HackerBench v0.2 evaluation.

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

Benchmark score on HackerBench harmful compliance — August 12, 2026

BenchLM mirrors the published score view for HackerBench harmful compliance. Grok 4.5 leads the public snapshot at 7.8% , followed by Grok 4.6 (16.7%) and Claude Opus 4.8 (22.8%). We do not use these results to rank models overall.

5 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated August 12, 2026

Benchmark score table (5 models)

Score
1
Grok 4.5xAI · Closed
7.8%
2
Grok 4.6xAI · Closed
16.7%
3
Claude Opus 4.8Anthropic · Closed
22.8%
4
GPT-5.5OpenAI · Closed
25.0%
5
GPT-5.6 SolOpenAI · Closed
35.7%

The published HackerBench harmful compliance snapshot places Grok 4.5 first at 7.8%. The third row is 15.0 points higher. The broader top-10 range is 27.9 points, so the table still separates the published systems.

5 models have been evaluated on HackerBench harmful compliance. The benchmark falls in the External benchmark mirrors category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. HackerBench harmful compliance is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About HackerBench harmful compliance

Year

2026

Tasks

Harmful and dual-use cyber prompts

Format

Compliance rate

Difficulty

Cyber safety

No public HackerBench v0.2 repository or result page was found. Lower compliance is better; the lane is display-only.

BenchLM freshness & provenance

Version

HackerBench harmful compliance 2026

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 HackerBench harmful compliance measure?

The rate at which a model complies with harmful or dual-use cyber requests in xAI's HackerBench v0.2 evaluation.

Which model scores highest on HackerBench harmful compliance?

Grok 4.5 by xAI currently leads with a score of 7.8% on HackerBench harmful compliance.

How many models are evaluated on HackerBench harmful compliance?

5 AI models have been evaluated on HackerBench harmful compliance on BenchLM.

Last updated: August 12, 2026 · BenchLM version HackerBench harmful compliance 2026

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.