← OpenAI · GPT Individual model view OpenAI · GPT

gpt-4o-mini

Appears in 6 benchmarksMean lean (confidence − pass rate): +0.215/6 benchmarks lean overconfident (prospective probe)

Positioning spread: every benchmark, one model

0.000.250.500.751.00SQuAD (factual recall)MMLU-Pro (knowledge)LegalBench (legal reasoning)MathBench (competition math)OmniMath (advanced math)SciCode (scientific code)performanceconfidence (red gap = overconfident)
Performance vs. confidence for gpt-4o-mini, per benchmark (prospective probe).
BenchmarkTask accConfidenceF₁Leans
SQuAD (factual recall)0.450.710.69+0.25 overconfident
MMLU-Pro (knowledge)0.400.870.59+0.47 overconfident
LegalBench (legal reasoning)0.850.450.60-0.39 cautious
MathBench (competition math)0.781.000.88+0.22 overconfident
OmniMath (advanced math)0.300.960.47+0.66 overconfident
SciCode (scientific code)0.370.390.56+0.02 calibrated

In the full cloud

-3-3-2-2-1-100112233Performance z-score within benchmark/probe →Confidence z-score →
gpt-4o-mini conditions all other model/condition points equal relative confidence and pass rate

Pairwise signal: pairs involving gpt-4o-mini

Match accuracy controls for the performance base-rate gap
gpt-4o-mini pairs
18/ 171
gpt-4o-mini mean tau
+0.007
All-pairs mean
+0.037
gpt-4o-mini p<0.05
3(16.7%)
-1.0-0.50.00.51.0Pair signal: do confidence gaps rank performance gaps? (Kendall tau-b)
all model pairs (observed) base-rate-matched null calibration-preserving null gpt-4o-mini pair (filled = p<0.05) gpt-4o-mini mean all-pairs mean

The four metacognitive outcomes

No curated cases for this selection yet — outcome-matrix extraction currently covers a sample of MMLU-Pro trials.
vs gpt-4o →vs gpt-5.2 → Compare with anything →