RLTasksby ulam.ai
Models / Google

Google / MODEL EVIDENCE

Gemini 2.5 Pro (Jun 2025)

22 benchmarks · 33 observations · 23 sources · released 2025-06-05

Scores at a glance

22 benchmarks · large dot: Gemini 2.5 Pro (Jun 2025); small dots: closest models where reported

GPQA Diamond ↗85.29%
9.2896.4

Gemini 2.5 Pro (Jun 2025) · GPQA diamond · Best score (across scorers) · % ↗

SWE-bench Verified ↗57.56%
30.9983.47

Gemini 2.5 Pro (Jun 2025) · SWE-Bench verified · Best score (across scorers) · % ↗

GSO ↗3.92%
026.5

Gemini 2.5 Pro (Jun 2025) · GSO-Bench · Score OPT@1 · % ↗

Terminal-Bench 2.0 ↗32.64%
3.0784.72

Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % ↗

7 more reported results

26.07% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

19.55% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

16.4% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

32.6% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

26.1% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

19.6% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

16.4% · Gemini 2.5 Pro (Jun 2025) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

CritPt ↗2%
032.29

Gemini 2.5 Pro (Jun 2025) · CritPt · Accuracy · % ↗

GDPval ↗23.3%
9.949.7

Gemini 2.5 Pro (Jun 2025) · GDPval · Win Rate (%) · % ↗

LMCA ↗34.77%
2.7868.26

Gemini 2.5 Pro (Jun 2025) · LMCA · Score · % ↗

DTBench ↗82.4%
41.6598.93

Gemini 2.5 Pro (Jun 2025) · DTBench · Accuracy · % ↗

SciCode ↗46.3%
1.2766.9

Gemini 2.5 Pro (Jun 2025) · SciCode · Score · % ↗

Closest models by score

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

7.7 percentile-point gap

6 of 7 shared benchmark comparisons

o3 →OpenAI
9.0 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. Arena leaderboard — WebDev evaluation ↗

    Arena

  2. SciCode Benchmark | Artificial Analysis ↗

    Artificial Analysis

  3. SpatialViz-Bench: A Cognitively-Grounded Benchmark for Diagnosing Spatial Visualization in MLLMs ↗

    Wang, Siting, Pei, Minnan, Sun, Luoyang, Deng, Cheng, Li, Yuchen, Shao, Kun et al. · 2025

    All 9 authors

    Wang, Siting, Pei, Minnan, Sun, Luoyang, Deng, Cheng, Li, Yuchen, Shao, Kun, Tian, Zheng, Zhang, Haifeng, Wang, Jun

  4. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  5. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  6. ARC-AGI-1 ↗

    Epoch AI

  7. ARC-AGI-2 ↗

    Epoch AI

  8. CritPt v1 ↗

    Epoch AI

  9. FrontierMath Tier 4 (v1, old) ↗

    Epoch AI

  10. FrontierMath Tier 4 (v2) ↗

    Epoch AI

  11. FrontierMath Tiers 1-3 (v1, old) ↗

    Epoch AI

  12. FrontierMath Tiers 1-3 (v2) ↗

    Epoch AI

  13. GDPval ↗

    Epoch AI

  14. GPQA Diamond ↗

    Epoch AI

  15. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  16. SWE-bench Verified ↗

    Epoch AI

  17. Terminal-Bench 2.0 ↗

    Epoch AI

  18. AI benchmarks & capabilities ↗

    Epoch AI

  19. FutureSearch Evals ↗

    FutureSearch

  20. Writing benchmark — incorporating mandatory story elements ↗

    lechmazur

  21. GSO: Challenging Software Optimization Tasks for Evaluating SWE-Agents ↗

    Shetty, Manish, Jain, Naman, Liu, Jinjian, Kethanaboyina, Vijay, Sen, Koushik, Stoica, Ion · 2025

  22. WeirdML ↗

    Håvard Tveit Ihle

  23. Terminal-Bench 2.0 leaderboard ↗

    Terminal-Bench Team

Names used in source tables

Gemini 2.5 Pro (Jun 2025)