Verdict
phi-3-medium 14B leads 7–1 across 8 shared benchmarks.
phi-3-medium 14B 7 · phi-3-mini 3.8B 1 · higher isn't always better — check the caveats below.
Reported scores — protocols may differ. How we compare
Reasoning: phi-3-medium 14B leads 2–1 across 3 shared reasoning benchmarks (largest gap: BBH, 75.2 vs 62.27 — single evaluator).
Commonsense Reasoning: phi-3-medium 14B leads 2–0 across 2 shared commonsense reasoning benchmarks (largest gap: Winogrande, 63 vs 41.6 — single evaluator).
phi-3-medium 14B leads 7–1 across 8 shared benchmarks. phi-3-medium 14B leads 7 benchmarks and phi-3-mini 3.8B leads 1. Higher isn't always better — see the integrity caveats on each benchmark.
phi-3-medium 14B and phi-3-mini 3.8B have 8 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.
phi-3-medium 14B leads commonsense reasoning 2–0. phi-3-medium 14B leads knowledge 2–0.