RLTasksby ulam.ai
Models / xAI

xAI / MODEL EVIDENCE

Grok 4

39 benchmarks · 60 observations · 46 sources

Scores at a glance

39 benchmarks · large dot: Grok 4; small dots: closest models where reported

BALROG ↗43.6%
3.768.3

Grok 4 · Balrog · Average progress · % ↗

Terminal-Bench 2.0 ↗27.19%
3.0784.72

Grok 4 · Terminal Bench · Accuracy mean · % ↗

5 more reported results

25.39% · Grok 4 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

23.15% · Grok 4 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

27.2% · Grok 4 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

25.4% · Grok 4 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

23.1% · Grok 4 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

GAIA-2 ↗8.8%
7.169.2

Grok-4 · GAIA-2 · Gaia2 full · Execution pass@1 · baseline ReAct · three runs ↗

7 more reported results

57.5% · Grok-4 · GAIA-2 · Gaia2 full · Search pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

9.4% · Grok-4 · GAIA-2 · Gaia2 full · Ambiguity pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

4.4% · Grok-4 · GAIA-2 · Gaia2 full · Adaptability pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

0% · Grok-4 · GAIA-2 · Gaia2 full · Time pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

15.6% · Grok-4 · GAIA-2 · Gaia2 full · Noise pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

14.4% · Grok-4 · GAIA-2 · Gaia2 full · Agent2Agent pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

15.7% · Grok-4 · GAIA-2 · Gaia2 full · Overall pass@1 · baseline ReAct · three runs · Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments ↗

Closest models by score

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

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. Aider LLM Leaderboards — Polyglot coding benchmark ↗

    Aider

  3. AIME 2025 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  4. Global-MMLU-Lite Benchmark | Artificial Analysis ↗

    Artificial Analysis

  5. LiveCodeBench Benchmark | Artificial Analysis ↗

    Artificial Analysis

  6. MATH-500 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  7. MMLU-Pro Benchmark | Artificial Analysis ↗

    Artificial Analysis

  8. SealQA: Raising the Bar for Reasoning in Search-Augmented Language Models — published result table ↗

    Pham, Thinh, Nguyen, Nguyen, Zunjare, Pratibha, Chen, Weiyuan, Tseng, Yu-Min, Vu, Tu · 2025

  9. TextQuests: How Good are LLMs at Text-Based Video Games? — published result table ↗

    Phan, Long, Mazeika, Mantas, Zou, Andy, Hendrycks, Dan · 2025

  10. Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments — published result table ↗

    Froger, Romain, Andrews, Pierre, Bettini, Matteo, Budhiraja, Amar, Cabral, Ricardo Silveira, Do, Virginie et al. · 2026

    All 24 authors

    Froger, Romain, Andrews, Pierre, Bettini, Matteo, Budhiraja, Amar, Cabral, Ricardo Silveira, Do, Virginie, Garreau, Emilien, Gaya, Jean-Baptiste, Laurençon, Hugo, Lecanu, Maxime, Malkan, Kunal, Mekala, Dheeraj, Ménard, Pierre, Bertran, Gerard Moreno-Torres, Piterbarg, Ulyana, Plekhanov, Mikhail, Rita, Mathieu, Rusakov, Andrey, Vorotilov, Vladislav, Wang, Mengjue, Yu, Ian, Benhalloum, Amine, Mialon, Grégoire, Scialom, Thomas

  11. BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games ↗

    Paglieri, Davide, Cupiał, Bartłomiej, Coward, Samuel, Piterbarg, Ulyana, Wolczyk, Maciej, Khan, Akbir et al. · 2024

    All 13 authors

    Paglieri, Davide, Cupiał, Bartłomiej, Coward, Samuel, Piterbarg, Ulyana, Wolczyk, Maciej, Khan, Akbir, Pignatelli, Eduardo, Kuciński, Łukasz, Pinto, Lerrel, Fergus, Rob, Foerster, Jakob Nicolaus, Parker-Holder, Jack, Rocktäschel, Tim

  12. Grok 4 Model Card ↗

    xAI · 2025

  13. FutureSearch Evals ↗

    FutureSearch

  14. ARC-AGI-1 ↗

    Epoch AI

  15. ARC-AGI-2 ↗

    Epoch AI

  16. FrontierMath Tier 4 (v1, old) ↗

    Epoch AI

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

    Epoch AI

  18. GDPval ↗

    Epoch AI

  19. GPQA Diamond ↗

    Epoch AI

  20. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  21. Terminal-Bench 2.0 ↗

    Epoch AI

  22. AI benchmarks & capabilities ↗

    Epoch AI

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

    Fiction.live · 2025

  24. GeoBench — GeoGuessr model benchmark ↗

    GeoBench

  25. Writing benchmark — incorporating mandatory story elements ↗

    lechmazur

  26. WeirdML ↗

    Håvard Tveit Ihle

  27. AI Model Benchmarks — LM Council ↗

    LM Council

  28. Measuring AI Ability to Complete Long Software Tasks ↗

    METR

  29. Terminal-Bench 2.0 leaderboard ↗

    Terminal-Bench Team

  30. AIME — evaluation methodology and results ↗

    Vals AI · 2026

  31. CaseLaw — evaluation methodology and results ↗

    Vals AI · 2026

  32. CorpFin — evaluation methodology and results ↗

    Vals AI · 2026

  33. GPQA Diamond — evaluation methodology and results ↗

    Vals AI · 2026

  34. LiveCodeBench — evaluation methodology and results ↗

    Vals AI · 2026

  35. LegalBench — evaluation methodology and results ↗

    Vals AI · 2026

  36. MATH 500 — evaluation methodology and results ↗

    Vals AI · 2026

  37. MedCode — evaluation methodology and results ↗

    Vals AI · 2026

  38. MedQA — evaluation methodology and results ↗

    Vals AI · 2026

  39. MedScribe — evaluation methodology and results ↗

    Vals AI · 2026

  40. MGSM — evaluation methodology and results ↗

    Vals AI · 2026

  41. MMLU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  42. MMMU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  43. MortgageTax — evaluation methodology and results ↗

    Vals AI · 2026

  44. SAGE — evaluation methodology and results ↗

    Vals AI · 2026

  45. SWE-bench Verified — evaluation methodology and results ↗

    Vals AI · 2026

  46. TaxEval — evaluation methodology and results ↗

    Vals AI · 2026

Release dates reported by sources

2025-07-09 · 2025-07-10

Names used in source tables

Grok 4 · Grok 4 (Mini-SWE-agent) · Grok-4 · grok-4 (high)