Skip to content

Benchmark

OpenAI Internal Design Tasks

Scores are percentages. Higher is better.

Models measured
3
Scores recorded
3
Sources
1
IQ weight
No weight published

Score distribution

44.4%
Mean
47.4%
Median
6.2%
Std dev
35.8%
10th percentile
50%
90th percentile
35.8%
Min
35.8%50%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
1GPT-6 AstraOpenAI internal benchmark· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.50%
2GPT-5.6 SolOpenAI internal benchmark· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.47.4%
3Claude Fable 5OpenAI internal benchmark· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.35.8%

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.