RLTasksby ulam.ai
Models / OpenAI

OpenAI / MODEL EVIDENCE

GPT-5 mini

49 benchmarks · 78 observations · 54 sources · released 2025-08-07

Scores at a glance

49 benchmarks · large dot: GPT-5 mini; small dots: closest models where reported

MATH Level 5 ↗97.85%
97.8597.85

GPT-5 mini (high) · MATH level 5 · Best score (across scorers) · % ↗

1 more reported result

96.79% · GPT-5 mini (medium) · MATH level 5 · Best score (across scorers) · % · Epoch AI · MATH level 5 ↗

Terminal-Bench 2.0 ↗34.83%
3.0784.72

GPT-5 mini (unknown thinking) · Terminal Bench · Accuracy mean · % ↗

9 more reported results

31.85% · GPT-5 mini (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

29.21% · GPT-5 mini (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

24.04% · GPT-5 mini (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

22.25% · GPT-5 mini (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

34.8% · GPT-5 mini (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

31.9% · GPT-5 mini (medium) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

29.2% · GPT-5 mini (medium) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

24% · GPT-5 mini (medium) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

22.2% · GPT-5 mini (medium) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

VPCT ↗40.2%
40.240.2

GPT-5 mini (medium) · VPCT · Correct · % ↗

1 more reported result

39% · GPT-5 mini (high) · VPCT · Correct · % · Epoch AI · VPCT ↗

LMCA ↗34.2%
2.7868.26

GPT-5 mini (high) · LMCA · Score · % ↗

DTBench ↗80.53%
41.6598.93

GPT-5 mini (high) · DTBench · Accuracy · % ↗

EnigmaEval ↗8.19%
0.3839.28

GPT-5 mini (unknown thinking) · EnigmaEval · Accuracy · % ↗

EconBench ↗0.0536 error points
0.050.05

gpt-5-mini · EconBench · EconBench prototype · risk-consistency error · published prototype observations ↗

Closest models by score

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

o3 →OpenAI
12.9 percentile-point gap

16 of 21 shared benchmark comparisons

14.1 percentile-point gap

14 of 21 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. AIME 2025 Benchmark | Artificial Analysis ↗

    Artificial Analysis

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

    Artificial Analysis

  3. LiveCodeBench Benchmark | Artificial Analysis ↗

    Artificial Analysis

  4. SciCode Benchmark | Artificial Analysis ↗

    Artificial Analysis

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

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

  6. CMT-Benchmark: A Benchmark for Condensed Matter Theory Built by Expert Researchers — published result table ↗

    Pan, Haining, Roggeveen, James V., Berg, Erez, Carrasquilla, Juan, Chowdhury, Debanjan, Ganguli, Surya et al. · 2025

    All 19 authors

    Pan, Haining, Roggeveen, James V., Berg, Erez, Carrasquilla, Juan, Chowdhury, Debanjan, Ganguli, Surya, Ghimenti, Federico, Hasik, Juraj, Hunt, Henry, Jiang, Hong-Chen, Kamb, Mason, Kao, Ying-Jer, Khatami, Ehsan, Lawler, Michael J., Luo, Di, Neupert, Titus, Qi, Xiaoliang, Brenner, Michael P., Kim, Eun-Ah

  7. OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning — published result table ↗

    Opsahl-Ong, Krista, Singhvi, Arnav, Collins, Jasmine, Zhou, Ivan, Wang, Cindy, Baheti, Ashutosh et al. · 2026

    All 13 authors

    Opsahl-Ong, Krista, Singhvi, Arnav, Collins, Jasmine, Zhou, Ivan, Wang, Cindy, Baheti, Ashutosh, Oertell, Owen, Portes, Jacob, Havens, Sam, Elsen, Erich, Bendersky, Michael, Zaharia, Matei, Chen, Xing

  8. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  9. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  10. EconBench: Economic choice by language models — gpt-5-mini prototype metrics ↗

    EconBench project · 2026

  11. Algotune — methodology and published results ↗

    Epoch AI · 2026

  12. ARC-AGI-1 ↗

    Epoch AI

  13. ARC-AGI-2 ↗

    Epoch AI

  14. Chess Puzzles ↗

    Epoch AI

  15. CritPt v1 ↗

    Epoch AI

  16. FrontierMath Tier 4 (v1, old) ↗

    Epoch AI

  17. FrontierMath Tier 4 (v2) ↗

    Epoch AI

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

    Epoch AI

  19. FrontierMath Tiers 1-3 (v2) ↗

    Epoch AI

  20. GPQA Diamond ↗

    Epoch AI

  21. Humanity's Last Exam ↗

    Epoch AI

  22. MATH Level 5 ↗

    Epoch AI

  23. Mystery Game Puzzles ↗

    Epoch AI

  24. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  25. SimpleQA Verified ↗

    Epoch AI

  26. SWE-bench Verified ↗

    Epoch AI

  27. Terminal-Bench 2.0 ↗

    Epoch AI

  28. Capabilities & benchmarking — retained external evaluation records ↗

    Epoch AI · 2026

  29. Vending Bench 2 — methodology and published results ↗

    Epoch AI · 2026

  30. AI benchmarks & capabilities ↗

    Epoch AI

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

    Fiction.live · 2025

  32. Writing benchmark — incorporating mandatory story elements ↗

    lechmazur

  33. WeirdML ↗

    Håvard Tveit Ihle

  34. EnigmaEval ↗

    Scale Labs / SEAL

  35. Terminal-Bench 2.0 leaderboard ↗

    Terminal-Bench Team

  36. AIME — evaluation methodology and results ↗

    Vals AI · 2026

  37. CaseLaw — evaluation methodology and results ↗

    Vals AI · 2026

  38. CorpFin — evaluation methodology and results ↗

    Vals AI · 2026

  39. GPQA Diamond — evaluation methodology and results ↗

    Vals AI · 2026

  40. LiveCodeBench — evaluation methodology and results ↗

    Vals AI · 2026

  41. LegalBench — evaluation methodology and results ↗

    Vals AI · 2026

  42. MATH 500 — evaluation methodology and results ↗

    Vals AI · 2026

  43. MedCode — evaluation methodology and results ↗

    Vals AI · 2026

  44. MedQA — evaluation methodology and results ↗

    Vals AI · 2026

  45. MedScribe — evaluation methodology and results ↗

    Vals AI · 2026

  46. MGSM — evaluation methodology and results ↗

    Vals AI · 2026

  47. MMLU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  48. MMMU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  49. MortgageTax — evaluation methodology and results ↗

    Vals AI · 2026

  50. ProofBench — evaluation methodology and results ↗

    Vals AI · 2026

  51. SAGE — evaluation methodology and results ↗

    Vals AI · 2026

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

    Vals AI · 2026

  53. TaxEval — evaluation methodology and results ↗

    Vals AI · 2026

  54. Vibe Code Bench — evaluation methodology and results ↗

    Vals AI · 2026

Names used in source tables

GPT 5 Mini · GPT 5 Mini (Mini-SWE-agent) · GPT 5 Mini (OpenHands) · GPT-5 Mini · GPT-5 Mini (high) · GPT-5 mini · GPT-5 mini (High) · GPT-5 mini (high) · GPT-5 mini (low) · GPT-5 mini (medium) · GPT-5 mini (unknown thinking) · GPT-5-mini · gpt-5-mini