Benchmark
PerceptionBench · Perception-related hallucination accuracy
Factuality benchmarks measure how often models get facts right. Scores are percentages. Higher is better.
- Models and systems measured
- 15
- Scores recorded
- 15
- Sources
- 1
- IQ weight
- 0.39
Score distribution
- 35.8%
- Mean
- 34.7%
- Median
- 7.6%
- Std dev
- 26.9%
- 10th percentile
- 45%
- 90th percentile
- 26.9%
- Min
This benchmark has a weight of 0.39 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 system | Score |
|---|---|---|
| 1 | Gemini 3.5 Flashbenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · high reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 50.6% |
| 2 | Claude Fable 5benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · max reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 45% |
| 3 | Grok 4.5benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 44.7% |
| 4 | Kimi K3benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · max reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 41.7% |
| 5 | Kimi K2.6benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 41% |
| 6 | Gemini 3.1 Pro Previewbenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · high reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 40.6% |
| 7 | Claude Opus 4.8benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · max reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 38.7% |
| 8 | GPT 5.5benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · xhigh reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 34.7% |
| 9 | MiniMax M3benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 29.9% |
| 10 | Qwen3.7 Plusbenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 29.5% |
| 11 | GPT-5.6 Solbenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · max reasoning · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 26.9% |
| 12 | Qwen3.5 397B A17Bbenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 26.9% |
Not ranked
The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.
| Model or system | Score |
|---|---|
| Gemma 4 31B (PerceptionBench label)benchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · thinking mode enabled · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 32.1% |
| GLM 5V Turbobenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 28% |
| GLM-4.6Vbenchmark-owner evaluation · 3,000 verified questions total · open-ended short answers · unified prompts · highest available reasoning budget; level not specified · main-table result; run count not specified · GPT-oss-120B judge· Reported by Moonshot AI / PerceptionBench authors· via BenchBoard docs | 26.9% |