Gemini 2.5 Pro (Jun 2025) vs Grok 4

Verdict

Evenly split 7–7 across 14 shared benchmarks.

Gemini 2.5 Pro (Jun 2025) 7 · Grok 4 7 · 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)Grok 4

83.1
AAider polyglot
79.6
41
BARC-AGI
66.67
4.86
AARC-AGI-2
15.97
15.82
AChess Puzzles
24.24
42.8
BDeepResearch Bench
47.3
91.7
AFiction.LiveBench
94.4
23.3
AGDPval
21.1
80.39
BGPQA diamond
82.67
83.8
BLech Mazur Writing
76.9
55.44
BMETR Time Horizons
66.58
84.71
BOTIS Mock AIME 2024-2025
83.98
54.88
ASimpleBench
52.6
32.64
BTerminal Bench
27.2
54.03
AWeirdML
45.73

Capabilities

Reasoning

Reasoning: Grok 4 leads 3–1 across 4 shared reasoning benchmarks (largest gap: ARC-AGI, 66.67 vs 41 — protocols undisclosed).

2 reproduced6 unverified

Coding

Coding: Gemini 2.5 Pro (Jun 2025) leads 2–0 across 2 shared coding benchmarks (largest gap: Terminal Bench, 32.64 vs 27.2 — single evaluator).

4 unverified

A higher score does not always mean a better model. Each number is task performance under a disclosed harness. Of the 28 measurements across 14 shared benchmarks, 28 single-evaluator or undisclosed-protocol reports; 21% are independently reproduced. Check any benchmark's integrity grade before reading a win as decisive.

Frequently asked questions

Evenly split 7–7 across 14 shared benchmarks. Gemini 2.5 Pro (Jun 2025) leads 7 benchmarks and Grok 4 leads 7. Higher isn't always better — see the integrity caveats on each benchmark.

Gemini 2.5 Pro (Jun 2025) and Grok 4 have 14 benchmarks in common in our data — those are the only rows where a direct, apples-to-apples comparison is drawn.

Grok 4 leads reasoning 3–1. Gemini 2.5 Pro (Jun 2025) leads coding 2–0.

Follow the record

These models

Same tier