Gemini 2.5 Pro (Jun 2025) vs o3

Verdict

o3 leads 13–7 across 20 shared benchmarks.

Gemini 2.5 Pro (Jun 2025) 7 · o3 13 · 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 (Jun 2025)o3

83.1
AAider polyglot
81.3
41
BARC-AGI
60.83
4.86
AARC-AGI-2
6.53
15.82
AChess Puzzles
33.71
42.8
BDeepResearch Bench
45.2
70.67
ADTBench
74.67
91.7
AFiction.LiveBench
88.9
24.56
AFrontierMath-Tiers-1-3-v2-Private
33.33
23.3
AGDPval
30.8
80.39
BGPQA diamond
75.76
3.92
AGSO-Bench
8.8
17.69
AHLE
16.3
83.8
BLech Mazur Writing
83.9
40.91
ALMCA
46.71
55.44
BMETR Time Horizons
65.44
84.71
BOTIS Mock AIME 2024-2025
84.43
54.88
ASimpleBench
43.72
57.56
BSWE-Bench verified
62.32
19.6
AVPCT
28
54.03
AWeirdML
52.42

Capabilities

Reasoning

Reasoning: Gemini 2.5 Pro (Jun 2025) leads 3–2 across 5 shared reasoning benchmarks (largest gap: ARC-AGI, 60.83 vs 41 — single evaluator).

2 reproduced8 unverified

Unknown

Unknown: o3 leads 3–0 across 3 shared unknown benchmarks (largest gap: METR Time Horizons, 65.44 vs 55.44 — single evaluator).

6 unverified

A higher score does not always mean a better model. Each number is task performance under a disclosed harness. Of the 40 measurements across 20 shared benchmarks, 40 single-evaluator or undisclosed-protocol reports; 20% are independently reproduced. Check any benchmark's integrity grade before reading a win as decisive.

Frequently asked questions

o3 leads 13–7 across 20 shared benchmarks. Gemini 2.5 Pro (Jun 2025) leads 7 benchmarks and o3 leads 13. Higher isn't always better — see the integrity caveats on each benchmark.

Gemini 2.5 Pro (Jun 2025) and o3 have 20 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.

o3 leads unknown 3–0. Gemini 2.5 Pro (Jun 2025) leads reasoning 3–2.

Follow the record

These models

Same tier