Skip to content

Benchmark

MentalHealthBench · Overall

Scores are percentages. Higher is better.

Models and systems measured
16
Scores recorded
16
Sources
1
IQ weight
0.85

Score distribution

44.7%
Mean
46.4%
Median
8.6%
Std dev
32.1%
10th percentile
57.3%
90th percentile
29.5%
Min
29.5%57.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.85 for chat models, measured over 11 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
1GPT-6.1 SolGPT-6.1 Sol system card · max reasoning · four-response task mean · 1,215 tasks· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.57.9%
2GPT-6 AstraOpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.57.3%
3GPT-6 SolOpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.53.9%
4Claude Opus 5.5OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs52.4%
5GPT-6 LunaOpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.50.2%
6Muse Spark 1.3OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs48.6%
7Claude Fable 5.1OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs46.4%
8Claude Sonnet 5OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs44.5%
9Claude Haiku 4.5OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs41.7%
10Grok 4.7OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs41.3%
11Gemini 3.8 FlashOpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs35.5%
12Gemini 3.1 Pro PreviewOpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs32.1%

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Gemini 2.5 ProOpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs29.5%
GPT-4o (March 2025)OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.32.1%
GPT-5.6 Luna (Aug 2026)OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.44.9%
GPT-5.6 Sol (Aug 2026)OpenAI evaluation · 95% CI available on source page· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.47%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.