Gemini 2.5 Pro (Mar 2025) vs o4-mini

Verdict

Gemini 2.5 Pro (Mar 2025) leads 7–4 across 11 shared benchmarks.

Gemini 2.5 Pro (Mar 2025) 7 · o4-mini 4 · higher isn't always better — check the caveats below.

Per-benchmark head-to-head

Reported scores — protocols may differ. How we compare

Sort

Gemini 2.5 Pro (Mar 2025)o4-mini

72.9
AAider polyglot
72
33
BARC-AGI
58.7
64
ACadEval
62
66.7
AFiction.LiveBench
77.8
81
AGeoBench
64
78.45
BGPQA diamond
72.81
14.03
AHLE
13.95
80.5
BLech Mazur Writing
75
95.56
BMATH level 5
97.83
41.92
ASimpleBench
26.44
22
AVPCT
36.25

Capabilities

Reasoning

Reasoning: Gemini 2.5 Pro (Mar 2025) leads 3–1 across 4 shared reasoning benchmarks (largest gap: ARC-AGI, 58.7 vs 33 — single evaluator).

2 reproduced6 unverified

A higher score does not always mean a better model. Each number is task performance under a disclosed harness. Of the 22 measurements across 11 shared benchmarks, 22 single-evaluator or undisclosed-protocol reports; 18% are independently reproduced. Check any benchmark's integrity grade before reading a win as decisive.

Frequently asked questions

Gemini 2.5 Pro (Mar 2025) leads 7–4 across 11 shared benchmarks. Gemini 2.5 Pro (Mar 2025) leads 7 benchmarks and o4-mini leads 4. Higher isn't always better — see the integrity caveats on each benchmark.

Gemini 2.5 Pro (Mar 2025) and o4-mini have 11 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.

Gemini 2.5 Pro (Mar 2025) leads reasoning 3–1.

Follow the record

These models

Same tier

Frequently asked questions

Gemini 2.5 Pro (Mar 2025) and o4-mini have 11 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.

Gemini 2.5 Pro (Mar 2025) leads reasoning 3–1.