Skip to content

Benchmark

PerceptionBench · Counting accuracy

Scores are percentages. Higher is better.

Models and systems measured
15
Scores recorded
15
Sources
1
IQ weight
0.96

Score distribution

46.3%
Mean
45.2%
Median
10.1%
Std dev
31.8%
10th percentile
59.7%
90th percentile
30.3%
Min
30.3%62.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.96 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-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 docs62.4%
2Kimi 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.59.7%
3Gemini 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 docs56.9%
4GPT 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 docs55.8%
5Qwen3.7 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 docs53.3%
6Claude 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 docs52.9%
7Qwen3.5 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 docs49.1%
8Kimi 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.45.2%
9Claude 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 docs44.2%
10Gemini 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 docs43.6%
11Grok 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 docs35.2%
12MiniMax 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 docs30.3%

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
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 docs33.9%
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 docs40%
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 docs31.8%

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.