benchgap
Other · model

Phi-3.5-mini benchmark scores

As of 2026-10-29, Phi-3.5-mini (Other) has measured scores on 5 benchmarks and estimated scores on 10 more.

BenchmarkScoreSource
AA-GPQA Diamond28.1%estimated ± 6.3 pp, medium confidence
AA-HLE4.1%estimated ± 5.0 pp, high confidence
CritPt0.0%estimated ± 3.4 pp, medium confidence
GPQA Diamond31.0%estimated ± 1.2 pp, high confidence
MMLU-Pro47.4%measured
Artificial Analysis Intelligence Index1.1%estimated ± 2.1 pp, medium confidence
BBH69.0%measured
GPQA30.4%measured
MMLU69.0%measured
AA-Omniscience Accuracy12.9%estimated ± 1.5 pp, medium confidence
MMMLU55.4%measured
C-Eval69.6%estimated ± 1.2 pp, medium confidence
MMLU-Redux69.2%estimated ± 0.6 pp, medium confidence
SuperGPQA27.1%estimated ± 12.1 pp, low confidence
SimpleQA0.0%estimated ± 14.6 pp, low confidence