Verdict
Claude Sonnet 4.5 leads 10–8 across 18 shared benchmarks.
Claude Sonnet 4.5 10 · GPT-5 mini 8 · higher isn't always better — check the caveats below.
Reported scores — protocols may differ. How we compare
Reasoning: Claude Sonnet 4.5 leads 3–1 across 4 shared reasoning benchmarks (largest gap: GPQA diamond, 76.43 vs 66.67 — single evaluator).
Unknown: Claude Sonnet 4.5 leads 4–0 across 4 shared unknown benchmarks (largest gap: ProofBench, 19 vs 9 — single evaluator).
Math: GPT-5 mini leads 3–0 across 3 shared math benchmarks (largest gap: FrontierMath-Tier-4-v2-Private, 12.2 vs 2.44 — single evaluator).
Claude Sonnet 4.5 leads 10–8 across 18 shared benchmarks. Claude Sonnet 4.5 leads 10 benchmarks and GPT-5 mini leads 8. Higher isn't always better — see the integrity caveats on each benchmark.
Claude Sonnet 4.5 and GPT-5 mini have 18 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.
Claude Sonnet 4.5 leads unknown 4–0. GPT-5 mini leads math 3–0.