Verdict
Qwen2.5-72B leads 6–2 across 8 shared benchmarks.
phi-3-medium 14B 2 · Qwen2.5-72B 6 · higher isn't always better — check the caveats below.
Reported scores — protocols may differ. How we compare
Reasoning: Qwen2.5-72B leads 2–1 across 3 shared reasoning benchmarks (largest gap: GPQA diamond, 32.2 vs 3.45 — single evaluator).
Commonsense Reasoning: Qwen2.5-72B leads 2–0 across 2 shared commonsense reasoning benchmarks (largest gap: HellaSwag, 79.73 vs 76.53 — single evaluator).
Qwen2.5-72B leads 6–2 across 8 shared benchmarks. phi-3-medium 14B leads 2 benchmarks and Qwen2.5-72B leads 6. Higher isn't always better — see the integrity caveats on each benchmark.
phi-3-medium 14B and Qwen2.5-72B have 8 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.
Qwen2.5-72B leads commonsense reasoning 2–0. Qwen2.5-72B leads reasoning 2–1.