Verdict
GPT-5 leads 23–0 across 23 shared benchmarks.
GPT-5 23 · o4-mini 0 · higher isn't always better — check the caveats below.
Reported scores — protocols may differ. How we compare
Reasoning: GPT-5 leads 5–0 across 5 shared reasoning benchmarks (largest gap: SimpleQA Verified, 50.1 vs 19.6 — protocols undisclosed).
Unknown: GPT-5 leads 4–0 across 4 shared unknown benchmarks (largest gap: DTBench, 84.45 vs 62.67 — single evaluator).
Math: GPT-5 leads 3–0 across 3 shared math benchmarks (largest gap: FrontierMath-Tier-4-v2-Private, 21.95 vs 4.88 — single evaluator).
GPT-5 leads 23–0 across 23 shared benchmarks. GPT-5 leads 23 benchmarks and o4-mini leads 0. Higher isn't always better — see the integrity caveats on each benchmark.
GPT-5 and o4-mini have 23 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.
GPT-5 leads reasoning 5–0. GPT-5 leads unknown 4–0.