RLTasksby ulam.ai
Models / OpenAI

OpenAI / MODEL EVIDENCE

GPT-5.6 Sol

116 benchmarks · 200 observations · 103 sources · released 2026-07-09

Scores at a glance

116 benchmarks · large dot: GPT-5.6 Sol; small dots: closest models where reported

ITBench-AA ↗56.21%
35.2256.21

GPT-5.6 Sol (Max) · ITBench-AA · root-cause score ↗

DeepSWE v1.1 ↗73.0
66.977

GPT-5.6 Sol · DeepSWE v1.1 ↗

6 more reported results

72.7% · GPT-5.6 Sol · DeepSWE v1.1 · GPT-6 Astra launch report ↗

72.67% · GPT-5.6 Sol (max) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

70.73% · GPT-5.6 Sol (xhigh) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

69.4% · GPT-5.6 Sol (high) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

61.06% · GPT-5.6 Sol (medium) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

45.35% · GPT-5.6 Sol (low) · DeepSWE · Pass@1 · % · Epoch AI · DeepSWE ↗

CursorBench 4.0 ↗41.7%
41.757.8

GPT-5.6 Sol · CursorBench 4.0 ↗

5 more reported results

41.7% · GPT-5.6 Sol (max) · CursorBench · Score · % · Epoch AI · CursorBench ↗

37.7% · GPT-5.6 Sol (xhigh) · CursorBench · Score · % · Epoch AI · CursorBench ↗

35.7% · GPT-5.6 Sol (high) · CursorBench · Score · % · Epoch AI · CursorBench ↗

31.1% · GPT-5.6 Sol (medium) · CursorBench · Score · % · Epoch AI · CursorBench ↗

24.6% · GPT-5.6 Sol (low) · CursorBench · Score · % · Epoch AI · CursorBench ↗

HealthBench ↗60.5%
52.163.4

GPT-5.6 Sol · HealthBench · HealthBench Professional · length-adjusted ↗

ErdősBench ↗3.120 / 4
2.683.23

GPT-5.6 Sol xhigh · ErdősBench · mean progress / 4 · Codex ↗

MirrorCode ↗20%
8.8977.4

GPT-5.6 Sol (high) · MirrorCode · Best score (across scorers) · % ↗

BALROG ↗60%
3.768.3

GPT-5.6 Sol (max) · Balrog · Average progress · % ↗

GSO ↗76.47%
9.888.24

GPT-5.6 Sol (unknown thinking) · GSO-Bench · Score OPT@1 · % ↗

GBAEval ↗52.59%
52.5952.59

GPT-5.6 Sol (unknown thinking) · GBAEval · Overall score · % ↗

DTBench ↗96%
9696

GPT-5.6 Sol (pro, max) · DTBench · Accuracy · % ↗

5 more reported results

95.73% · GPT-5.6 Sol (xhigh) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

95.47% · GPT-5.6 Sol (max) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

95.47% · GPT-5.6 Sol (high) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

95.2% · GPT-5.6 Sol (medium) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

93.87% · GPT-5.6 Sol (low) · DTBench · Accuracy · % · Epoch AI · DTBench ↗

GDP.pdf ↗30.7%
30.730.7

GPT-5.6 Sol (unknown thinking) · GDP.pdf · GDP.pdf score · % ↗

1 more reported result

30.7% · GPT-5.6 Sol (max) · GDP.pdf · GDP.pdf score · % · Epoch AI · GDP.pdf ↗

Vals RSI Index ↗0.41 reference score
00.46

GPT-5.6 Sol (max) · Compression · reference-anchored score ↗

3 more reported results

0.36 reference score · GPT-5.6 Sol (max) · LM training · reference-anchored score · Vals RSI Index · LM training ↗

0.18 reference score · GPT-5.6 Sol (max) · Harness engineering · reference-anchored score · Vals RSI Index · Harness engineering ↗

0.00 reference score · GPT-5.6 Sol (max) · Post-training · reference-anchored score · Vals RSI Index · Post-training ↗

Fortress ↗20.14%
5.7874.39

gpt-5.6-sol · Scale Labs · Risk elicitation score ↗

BoxBench ↗74%
7277

GPT-5.6 Sol · BoxBench · weighted rubric accuracy · OpenAI · ChatGPT financial services launch ↗

Kimi Code Bench 2.0 ↗64.8%
64.276.9

GPT-5.6 Sol (max) · Kimi Code Bench 2.0 · score · Kimi K3 · official model-card comparison ↗

FrontierMath: Open Problems ↗2 credited problems
12

GPT-5.6 Sol · FrontierMath: Open Problems · FM Open Problems · first credited AI solution records · Epoch public solution registry · October 27, 2026 ↗

Closest models by score

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

17.7 percentile-point gap

43 of 58 shared benchmark comparisons

19.9 percentile-point gap

37 of 58 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 →

Reports featuring this model

DeepSeek V4.1 Flash 1017 · DeepSeek →

Claude Opus 5.5 · Anthropic →

GPT-6 Astra · OpenAI →

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. Blueprint-Bench 2 | Andon Labs ↗

    Andon Labs

  2. ARC Prize - Leaderboard ↗

    ARC Prize Foundation

  3. Arena leaderboard — WebDev evaluation ↗

    Arena

  4. AA-AnalystAgent Benchmark | Artificial Analysis ↗

    Artificial Analysis

  5. CritPt Benchmark | Artificial Analysis ↗

    Artificial Analysis

  6. CritPt v2 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  7. EnterpriseOps-Gym-AA Benchmark | Artificial Analysis ↗

    Artificial Analysis

  8. GPQA Diamond Benchmark | Artificial Analysis ↗

    Artificial Analysis

  9. ITBench-AA Benchmark | Artificial Analysis ↗

    Artificial Analysis

  10. MMMU-Pro Benchmark | Artificial Analysis ↗

    Artificial Analysis

  11. SciCode Benchmark | Artificial Analysis ↗

    Artificial Analysis

  12. 𝜏³-Banking Benchmark | Artificial Analysis ↗

    Artificial Analysis

  13. Terminal-Bench 2.1 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  14. Terminal-Bench Hard Benchmark | Artificial Analysis ↗

    Artificial Analysis

  15. TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts — published result table ↗

    Jaffe, Oliver, Sherburn, Dane · 2026

  16. 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

  17. FrontierCode Leaderboard ↗

    Cognition

  18. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  19. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  20. Cursor · CursorBench ↗

    Cursor

  21. DeepSWE ↗

    Datacurve

  22. APEX-Agents ↗

    Epoch AI

  23. ARC-AGI-2 ↗

    Epoch AI

  24. Chess Puzzles ↗

    Epoch AI

  25. CritPt v1 ↗

    Epoch AI

  26. Earthborne Rangers (EBR-bench) ↗

    Epoch AI

  27. FrontierMath Erdős ↗

    Epoch AI

  28. FrontierMath Tier 4 (v2) ↗

    Epoch AI

  29. FrontierMath Tiers 1-3 (v2) ↗

    Epoch AI

  30. Furniture Assembly Benchmark ↗

    Epoch AI

  31. GPQA Diamond ↗

    Epoch AI

  32. MirrorCode ↗

    Epoch AI

  33. Mystery Game Puzzles ↗

    Epoch AI

  34. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  35. PostTrainBench v1.1 ↗

    Epoch AI

  36. SimpleQA Verified ↗

    Epoch AI

  37. SimpleBench ↗

    Epoch AI

  38. Vending Bench 2 — methodology and published results ↗

    Epoch AI · 2026

  39. AI benchmarks & capabilities ↗

    Epoch AI

  40. FrontierMath: Open Problems — public problem and solution registry ↗

    Epoch AI · 2026

  41. FutureSearch Evals ↗

    FutureSearch

  42. GPT-5.6 Sol (unknown thinking) — GBA Eval ↗

    GBA Eval

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

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

  44. WeirdML ↗

    Håvard Tveit Ihle

  45. deepseek-ai/DeepSeek-V4.1-Flash-1017 · Hugging Face ↗

    DeepSeek AI

  46. Kimi K3: Open Frontier Intelligence — official model card ↗

    Moonshot AI · 2026

  47. DistressBench — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  48. DrugDiscoveryBench — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  49. EnigmaEval ↗

    Scale Labs / SEAL

  50. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  51. HiL-Bench (Human-in-Loop Benchmark) — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  52. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  53. Humanity's Sixth Sense — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  54. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  55. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  56. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  57. SWE-Bench Pro V2 — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  58. SWE-Bench Pro V2 — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  59. AI Model Benchmarks — LM Council ↗

    LM Council

  60. GPT-6 Astra: A new generation of intelligence ↗

    OpenAI

  61. Introducing ChatGPT for financial services ↗

    OpenAI · 2026

  62. OSWorld 2.0: Benchmarking computer-use agents on long-horizon real-world tasks ↗

    OSWorld Team

  63. GDP.pdf Benchmark | Surge AI ↗

    Surge AI

  64. Introducing Claude Opus 5.5 ↗

    Anthropic

  65. FrontierSWE — software engineering benchmark ↗

    FrontierSWE

  66. ErdősBench: research mathematics, scoped progress and proof audits ↗

    Ulam AI · 2026

  67. BioMysteryBench — evaluation methodology and results ↗

    Vals AI · 2026

  68. Code Migration v1 — evaluation methodology and results ↗

    Vals AI · 2026

  69. CorpFin — evaluation methodology and results ↗

    Vals AI · 2026

  70. CUA-bench — evaluation methodology and results ↗

    Vals AI · 2026

  71. CyberBench — evaluation methodology and results ↗

    Vals AI · 2026

  72. Excel Modeling Benchmark v1 — evaluation methodology and results ↗

    Vals AI · 2026

  73. Finance Agent v2 — evaluation methodology and results ↗

    Vals AI · 2026

  74. GPQA Diamond — evaluation methodology and results ↗

    Vals AI · 2026

  75. Harvey's Legal Agent Benchmark v1 — evaluation methodology and results ↗

    Vals AI · 2026

  76. IOI — evaluation methodology and results ↗

    Vals AI · 2026

  77. LiveCodeBench — evaluation methodology and results ↗

    Vals AI · 2026

  78. LegalBench — evaluation methodology and results ↗

    Vals AI · 2026

  79. Legal Research Bench — evaluation methodology and results ↗

    Vals AI · 2026

  80. MedCode — evaluation methodology and results ↗

    Vals AI · 2026

  81. MedScribe — evaluation methodology and results ↗

    Vals AI · 2026

  82. MMLU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  83. MMMU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  84. MortgageTax — evaluation methodology and results ↗

    Vals AI · 2026

  85. MysteryMechanism — evaluation methodology and results ↗

    Vals AI · 2026

  86. ProgramBench — evaluation methodology and results ↗

    Vals AI · 2026

  87. ProofBench — evaluation methodology and results ↗

    Vals AI · 2026

  88. Public Benefits Bench — evaluation methodology and results ↗

    Vals AI · 2026

  89. Vals RSI Index — evaluation methodology and results ↗

    Vals AI · 2026

  90. SAGE — evaluation methodology and results ↗

    Vals AI · 2026

  91. Vals Science Index — evaluation methodology and results ↗

    Vals AI · 2026

  92. SkillsBench — evaluation methodology and results ↗

    Vals AI · 2026

  93. SRE Bench — evaluation methodology and results ↗

    Vals AI · 2026

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

    Vals AI · 2026

  95. Tax Agent Bench — evaluation methodology and results ↗

    Vals AI · 2026

  96. TaxEval — evaluation methodology and results ↗

    Vals AI · 2026

  97. Terminal-Bench 4.0 — evaluation methodology and results ↗

    Vals AI · 2026

  98. Terminal-Bench-Science — evaluation methodology and results ↗

    Vals AI · 2026

  99. Time Horizon Index: KSP — evaluation methodology and results ↗

    Vals AI · 2026

  100. Vals Multimodal Index — evaluation methodology and results ↗

    Vals AI · 2026

  101. Vibe Code Bench 1-100 — evaluation methodology and results ↗

    Vals AI · 2026

  102. Vibe Code Bench — evaluation methodology and results ↗

    Vals AI · 2026

  103. Surface Evolver Bench ↗

    yhenon

Names used in source tables

GPT 5.6 Sol · GPT 5.6 Sol (Codex) max · GPT-5.6 Sol · GPT-5.6 Sol (Codex) xhigh · GPT-5.6 Sol (Max) · GPT-5.6 Sol (Mini-SWE-agent) · GPT-5.6 Sol (OpenHands) · GPT-5.6 Sol (Xhigh) · GPT-5.6 Sol (high) · GPT-5.6 Sol (low) · GPT-5.6 Sol (max) · GPT-5.6 Sol (medium) · GPT-5.6 Sol (pro, max) · GPT-5.6 Sol (pro, unknown thinking) · GPT-5.6 Sol (pro, xhigh) · GPT-5.6 Sol (unknown thinking) · GPT-5.6 Sol (xhigh) · GPT-5.6 Sol xhigh · GPT-5.6-Sol (max) · GPT-5.6-sol (Codex) xhigh · gpt-5.6 (sol) · gpt-5.6-sol · gpt-5.6-sol (max)