RLTasksby ulam.ai
Models / Google

Google / MODEL EVIDENCE

Gemini 2.5 Pro Preview (Mar 2025)

7 benchmarks · 7 observations · 7 sources · released 2025-03-31

Scores at a glance

7 benchmarks · large dot: Gemini 2.5 Pro Preview (Mar 2025); small dots: closest models where reported

MATH Level 5 ↗95.56%
3.2998.13

Gemini 2.5 Pro Preview (Mar 2025) · MATH level 5 · Best score (across scorers) · % ↗

ARC-AGI-1 ↗33%
098.5

Gemini 2.5 Pro Preview (Mar 2025) · ARC-AGI · Score · % ↗

CadEval ↗64%
1274

Gemini 2.5 Pro Preview (Mar 2025) · CadEval · Overall pass (%) · % ↗

SimpleBench ↗51.6%
10.781.9

Gemini 2.5 Pro Preview (Mar 2025) · SimpleBench · Score (AVG@5) · % ↗

VPCT ↗48%
3091

Gemini 2.5 Pro Preview (Mar 2025) · VPCT · Correct · % ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

o3 →OpenAI
10.5 percentile-point gap

7 of 7 shared benchmark comparisons

12.7 percentile-point gap

6 of 7 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. Aider code editing benchmark ↗

    Paul Gauthier, Aider contributors

  2. ARC Prize - Leaderboard ↗

    ARC Prize Foundation

  3. MATH Level 5 ↗

    Epoch AI

  4. SimpleBench ↗

    Epoch AI

  5. Capabilities & benchmarking — retained external evaluation records ↗

    Epoch AI · 2026

  6. Fiction.liveBench — March 14, 2025 benchmark release ↗

    Fiction.live · 2025

  7. CadEval Dashboard — run 20250422_095709 ↗

    willpatrick · 2025

Names used in source tables

Gemini 2.5 Pro Preview (Mar 2025)