openai/gpt-4o-mini
OpenAI
Consistency
How often it gives the same answer when asked again. Across 40 questions. How it's measured →
Alignment
How often its answer is the one most other models gave. Across 41 questions. How it's measured →
Hedging
How often it qualifies its answer or declines to give one. How it's measured →
Answers most like
Answers least like
Measured across the questions both models were asked, at least 20 of them. How it's measured →
Where this model stands apart
11 of the 41 scored questions where its answer wasn't the one most models gave.
10 runs · consistency 100
15 runs · consistency 77.8
10 runs · consistency 100
5 runs
10 runs · consistency 61.1
20 runs · consistency 40
20 runs · consistency 40
20 runs · consistency 40
20 runs · consistency 46.7
10 runs · consistency 100
10 runs · consistency 100
5 runs
10 runs · consistency 100
10 runs · consistency 40
25 runs · consistency 36.1
10 runs · consistency 100
10 runs
30 runs · consistency not scored
15 runs · consistency 50
15 runs · consistency 50
10 runs · consistency 100
20 runs · consistency 100
20 runs · consistency 100
20 runs · consistency 100
19 runs · consistency 100
10 runs · consistency 46.7
10 runs · consistency 40
20 runs · consistency 66.7
10 runs · consistency 40
10 runs · consistency 66.7
20 runs · consistency 22.2
5 runs
5 runs
10 runs · consistency 100