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

xAI factuality hallucination rate (Factuality hallucination rate)

The share of factuality-evaluation answers classified as hallucinations in the Grok 4.6 model card.

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

Benchmark score on Factuality hallucination rate — August 12, 2026

BenchLM mirrors the published score view for Factuality hallucination rate. Grok 4.5 leads the public snapshot at 0.98% , followed by GPT-5.5 (1.10%) and Grok 4.6 (1.70%). We do not use these results to rank models overall.

4 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated August 12, 2026

Benchmark score table (4 models)

Score
1
Grok 4.5xAI · Closed
0.98%
2
GPT-5.5OpenAI · Closed
1.10%
3
Grok 4.6xAI · Closed
1.70%
4
Claude Opus 4.8Anthropic · Closed
3.40%

The published Factuality hallucination rate snapshot places Grok 4.5 first at 0.98%. The third row is 0.72 points higher. The broader top-10 range is 2.42 points, so many of the published results sit in a relatively narrow band.

4 models have been evaluated on Factuality hallucination rate. 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. Factuality hallucination rate is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About Factuality hallucination rate

Year

2026

Tasks

Internal factuality questions

Format

Hallucination rate

Difficulty

Factual reliability

No public task set or matching result page was found. BenchLM preserves the exact card values as display-only internal-evaluation evidence.

BenchLM freshness & provenance

Version

Factuality hallucination rate 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 Factuality hallucination rate measure?

The share of factuality-evaluation answers classified as hallucinations in the Grok 4.6 model card.

Which model scores highest on Factuality hallucination rate?

Grok 4.5 by xAI currently leads with a score of 0.98% on Factuality hallucination rate.

How many models are evaluated on Factuality hallucination rate?

4 AI models have been evaluated on Factuality hallucination rate on BenchLM.

Compare Top Models on Factuality hallucination rate

Last updated: August 12, 2026 · BenchLM version Factuality hallucination rate 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.