Benchmark Almanac

SimpleBench

Claude Opus 5.5 88.4% leads since Sep 2026 Not in model cards

Models answer over 200 multiple-choice trick questions on spatial, temporal, and social reasoning that ordinary people solve. Scored by average accuracy over five runs.

By September 2026 Claude Opus 5.5 led at 88.4 percent, above the 83.7 percent baseline from nine non-expert people. A small open-ended variant lists three models.

Made by SimpleBench Team Released Oct 2024 Abstract puzzles and logic Public tasks plus a private test set Last model added Sep 2026 Scores refreshed 2 Nov 2026
Level
Everyday Most adults, without training
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Multiple choice
simple-bench/SimpleBench 220 stars Leaderboard simple-bench.com
104
Entries on its own leaderboard, 100 models we track

Claude Opus 5.5 leads with 88.4%

Best score per entry from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

104 entries. Scroll sideways for the rest.

  1. 88.4 Claude Opus 5.5
  2. 86.6 Claude Fable 5.1
  3. 86.5 GPT-6 Astra Pro
  4. 83.6 GPT-6 Astra
  5. 82.9 GPT-6.1 Sol
  6. 82.4 Gemini 3.8 Flash
  7. 81.9 Claude Fable 5
  8. 81.8 Muse Spark 1.3
  9. 80.6 Claude Opus 5
  10. 79.6 Gemini 3.1 Pro
  11. 76.9 GPT-5.5 Pro
  12. 76.7 Gemini 3.5 Flash
  13. 76.4 Gemini 3 Pro
  14. 75.9 Grok 4.6
  15. 75.9 Claude Sonnet 5.5
  16. 74.5 Muse Spark 1.2
  17. 73.1 GPT-6 Sol
  18. 71.7 GPT-5.6 Sol Pro
  19. 70.4 Qwen3.7-Max
  20. 70.0 Grok 4.5
  21. 69.0 GPT-5.5
  22. 67.6 Claude Opus 4.6
  23. 66.7 DeepSeek V4.1 Flash
  24. 66.2 GLM-5.3
  25. 64.8 Claude Opus 4.8
  26. 64.8 GPT-5.6 Sol
  27. 63.0 Qwen 3.6 Max (Preview)
  28. 62.5 Qwen 3.8 Max
  29. 62.4 Gemini 2.5 Pro
  30. 62.0 Claude Opus 4.5
  31. 61.7 Claude Opus 4.7
  32. 61.6 GPT-5 Pro
  33. 61.1 Gemini 3 Flash
  34. 61.1 DeepSeek-V4-Flash
  35. 60.7 Kimi K3
  36. 60.6 Claude Sonnet 5
  37. 60.5 Grok 4
  38. 60.2 Qwen 3.8 27B
  39. 60.0 Claude Opus 4.1
  40. 58.8 Claude Opus 4
  41. 58.8 GLM-5.2
  42. 57.9 Kimi K2.7 Code
  43. 57.4 GPT-5.2 Pro
  44. 56.7 GPT-5
  45. 56.0 Grok 4.1 Fast
  46. 55.1 GLM-5.1
  47. 55.1 Ox Alpha
  48. 54.3 Claude Sonnet 4.5
  49. 53.2 GPT-5.1
  50. 53.2 GLM-5
  51. 53.1 o3
  52. 52.6 DeepSeek-V3.2-Speciale
  53. 51.6 Gemini 2.5 Pro
  54. 50.9 DeepSeek-V4-Pro
  55. 50.0 Inkling
  56. 48.9 GPT-5.6 Terra
  57. 47.7 GLM-4.7
  58. 46.8 Kimi K2.5
  59. 46.8 GPT-5.6 Luna
  60. 46.4 Claude 3.7 Sonnet
  61. 46.3 DeepSeek V4 Flash (preview)
  62. 45.8 GPT-5.2
  63. 45.8 MiniMax-M3
  64. 45.5 Claude Sonnet 4
  65. 41.7 o1-preview
  66. 41.7 Nemotron 3 Ultra
  67. 41.4 Claude 3.5 Sonnet (October 2024)
  68. 41.2 Gemini 2.5 Flash
  69. 40.8 DeepSeek-R1
  70. 40.1 o1
  71. 40.0 DeepSeek-V3.1
  72. 39.6 Kimi K2 Thinking
  73. 38.7 o4-mini
  74. 36.1 Grok 3
  75. 35.2 Qwen 3.6 Flash
  76. 34.7 MiniMax-M2.1
  77. 34.5 GPT-4.5
  78. 31.1 Gemini-exp-1206
  79. 31.0 Qwen3-235B-A22B
  80. 30.9 DeepSeek-R1
  81. 30.7 Gemini 2.0 Flash Thinking
  82. 27.7 Llama 4 Maverick
  83. 27.5 Claude 3.5 Sonnet
  84. 27.2 DeepSeek-V3
  85. 27.1 Gemini 1.5 Pro
  86. 27.0 GPT-4.1
  87. 26.3 Kimi K2
  88. 25.1 GPT-4 Turbo
  89. 25.0 MiniMax-M2
  90. 23.5 Claude 3 Opus
  91. 23.0 Llama 3.1-405B
  92. 22.8 o3-mini
  93. 22.7 Grok-2
  94. 22.5 Mistral Large v2
  95. 22.1 gpt-oss-120b
  96. 20.4 Mistral Large 3
  97. 19.9 Llama 3.3 70B Instruct
  98. 18.9 Gemini 2.0 Flash
  99. 18.9 DeepSeek-V3
  100. 18.1 o1-mini
  101. 17.8 GPT-4o
  102. 17.4 Command R+
  103. 10.7 GPT-4o mini
  104. 8.00 GPT-3.5 Turbo
  1. 1 Claude Opus 5.5 88.4%
  2. 2 Claude Fable 5.1 86.6%
  3. 3 GPT-6 Astra Pro 86.5%
  4. 4 GPT-6 Astra 83.6%
  5. 5 GPT-6.1 Sol 82.9%
  6. 6 Gemini 3.8 Flash 82.4%
  7. 7 Claude Fable 5 81.9%
  8. 8 Muse Spark 1.3 81.8%
  9. 9 Claude Opus 5 80.6%
  10. 10 Gemini 3.1 Pro 79.6%
  11. 11 GPT-5.5 Pro 76.9%
  12. 12 Gemini 3.5 Flash 76.7%
  13. 13 Gemini 3 Pro 76.4%
  14. 14 Grok 4.6 75.9%
  15. 15 Claude Sonnet 5.5 75.9%
  16. 16 Muse Spark 1.2 74.5%
  17. 17 GPT-6 Sol 73.1%
  18. 18 GPT-5.6 Sol Pro xhigh 71.7%
  19. 19 Qwen3.7-Max 70.4%
  20. 20 Grok 4.5 70%
  21. 21 GPT-5.5 69%
  22. 22 Claude Opus 4.6 67.6%
  23. 23 DeepSeek V4.1 Flash 66.7%
  24. 24 GLM-5.3 66.2%
  25. 25 Claude Opus 4.8 64.8%
  26. 26 GPT-5.6 Sol xhigh 64.8%
  27. 27 Qwen 3.6 Max (Preview) 63%
  28. 28 Qwen 3.8 Max 62.5%
  29. 29 Gemini 2.5 Pro Jun 2025 62.4%
  30. 30 Claude Opus 4.5 62%
  31. 31 Claude Opus 4.7 61.7%
  32. 32 GPT-5 Pro 61.6%
  33. 33 Gemini 3 Flash 61.1%
  34. 34 DeepSeek-V4-Flash 61.1%
  35. 35 Kimi K3 max 60.7%
  36. 36 Claude Sonnet 5 60.6%
  37. 37 Grok 4 60.5%
  38. 38 Qwen 3.8 27B 60.2%
  39. 39 Claude Opus 4.1 60%
  40. 40 Claude Opus 4 58.8%
  41. 41 GLM-5.2 58.8%
  42. 42 Kimi K2.7 Code 57.9%
  43. 43 GPT-5.2 Pro xhigh 57.4%
  44. 44 GPT-5 high 56.7%
  45. 45 Grok 4.1 Fast 56%
  46. 46 GLM-5.1 55.1%
  47. 47 Ox Alpha 55.1%
  48. 48 Claude Sonnet 4.5 54.3%
  49. 49 GPT-5.1 high 53.2%
  50. 50 GLM-5 53.2%
  51. 51 o3 high 53.1%
  52. 52 DeepSeek-V3.2-Speciale 52.6%
  53. 53 Gemini 2.5 Pro Mar 2025 51.6%
  54. 54 DeepSeek-V4-Pro 50.9%
  55. 55 Inkling 50%
  56. 56 GPT-5.6 Terra xhigh 48.9%
  57. 57 GLM-4.7 47.7%
  58. 58 Kimi K2.5 46.8%
  59. 59 GPT-5.6 Luna xhigh 46.8%
  60. 60 Claude 3.7 Sonnet thinking 46.4%
  61. 61 DeepSeek V4 Flash (preview) 46.3%
  62. 62 GPT-5.2 high 45.8%
  63. 63 MiniMax-M3 45.8%
  64. 64 Claude Sonnet 4 thinking 45.5%
  65. 65 o1-preview 41.7%
  66. 66 Nemotron 3 Ultra 41.7%
  67. 67 Claude 3.5 Sonnet (October 2024) 41.4%
  68. 68 Gemini 2.5 Flash Sep 2025 41.2%
  69. 69 DeepSeek-R1 May 2025 40.8%
  70. 70 o1 high 40.1%
  71. 71 DeepSeek-V3.1 40%
  72. 72 Kimi K2 Thinking 39.6%
  73. 73 o4-mini high 38.7%
  74. 74 Grok 3 36.1%
  75. 75 Qwen 3.6 Flash 35.2%
  76. 76 MiniMax-M2.1 34.7%
  77. 77 GPT-4.5 34.5%
  78. 78 Gemini-exp-1206 31.1%
  79. 79 Qwen3-235B-A22B 31%
  80. 80 DeepSeek-R1 30.9%
  81. 81 Gemini 2.0 Flash Thinking Jan 2025 30.7%
  82. 82 Llama 4 Maverick 27.7%
  83. 83 Claude 3.5 Sonnet 27.5%
  84. 84 DeepSeek-V3 Mar 2025 27.2%
  85. 85 Gemini 1.5 Pro Sep 2024 27.1%
  86. 86 GPT-4.1 27%
  87. 87 Kimi K2 Jul 2025 26.3%
  88. 88 GPT-4 Turbo Apr 2024 25.1%
  89. 89 MiniMax-M2 25%
  90. 90 Claude 3 Opus 23.5%
  91. 91 Llama 3.1-405B 23%
  92. 92 o3-mini high 22.8%
  93. 93 Grok-2 Aug 2024 22.7%
  94. 94 Mistral Large v2 22.5%
  95. 95 gpt-oss-120b 22.1%
  96. 96 Mistral Large 3 20.4%
  97. 97 Llama 3.3 70B Instruct 19.9%
  98. 98 Gemini 2.0 Flash Dec 2024 18.9%
  99. 99 DeepSeek-V3 18.9%
  100. 100 o1-mini 18.1%
  101. 101 GPT-4o Aug 2024 17.8%
  102. 102 Command R+ 17.4%
  103. 103 GPT-4o mini 10.7%
  104. 104 GPT-3.5 Turbo Jun 2023 8.00%
#EntryScoreSettingDateSource
1 Claude Opus 5.5 88.4% Sep 2026 Official leaderboard
2 Claude Fable 5.1 86.6% Sep 2026 Official leaderboard
3 GPT-6 Astra Pro 86.5% Sep 2026 Official leaderboard
4 GPT-6 Astra 83.6% Sep 2026 Official leaderboard
5 GPT-6.1 Sol 82.9% Sep 2026 Official leaderboard
6 Gemini 3.8 Flash 82.4% Sep 2026 Official leaderboard
7 Claude Fable 5 81.9% Jun 2026 Official leaderboard
8 Muse Spark 1.3 81.8% Sep 2026 Official leaderboard
9 Claude Opus 5 80.6% Jul 2026 Official leaderboard
10 Gemini 3.1 Pro 79.6% Feb 2026 Official leaderboard
11 GPT-5.5 Pro 76.9% Apr 2026 Official leaderboard
12 Gemini 3.5 Flash 76.7% May 2026 Official leaderboard
13 Gemini 3 Pro 76.4% Nov 2025 Official leaderboard
14 Grok 4.6 75.9% Aug 2026 Official leaderboard
15 Claude Sonnet 5.5 75.9% Sep 2026 Official leaderboard
16 Muse Spark 1.2 74.5% Aug 2026 Official leaderboard
17 GPT-6 Sol 73.1% Sep 2026 Official leaderboard
18 GPT-5.6 Sol Pro 71.7% xhigh Jul 2026 Official leaderboard
19 Qwen3.7-Max 70.4% May 2026 Official leaderboard
20 Grok 4.5 70% Jul 2026 Official leaderboard
21 GPT-5.5 69% Apr 2026 Official leaderboard
22 Claude Opus 4.6 67.6% Feb 2026 Official leaderboard
23 DeepSeek V4.1 Flash 66.7% Sep 2026 Official leaderboard
24 GLM-5.3 66.2% Aug 2026 Official leaderboard
25 Claude Opus 4.8 64.8% May 2026 Official leaderboard
26 GPT-5.6 Sol 64.8% xhigh Jul 2026 Official leaderboard
27 Qwen 3.6 Max (Preview) 63% Apr 2026 Official leaderboard
28 Qwen 3.8 Max 62.5% Aug 2026 Official leaderboard
29 Gemini 2.5 Pro Jun 2025 62.4% Jun 2025 Official leaderboard
30 Claude Opus 4.5 62% Nov 2025 Official leaderboard
31 Claude Opus 4.7 61.7% Apr 2026 Official leaderboard
32 GPT-5 Pro 61.6% Oct 2025 Official leaderboard
33 Gemini 3 Flash 61.1% Dec 2025 Official leaderboard
34 DeepSeek-V4-Flash 61.1% Apr 2026 Official leaderboard
35 Kimi K3 60.7% max Jul 2026 Official leaderboard
36 Claude Sonnet 5 60.6% Jun 2026 Official leaderboard
37 Grok 4 60.5% Jul 2025 Official leaderboard
38 Qwen 3.8 27B 60.2% Aug 2026 Official leaderboard
39 Claude Opus 4.1 60% Aug 2025 Official leaderboard
40 Claude Opus 4 58.8% May 2025 Official leaderboard
41 GLM-5.2 58.8% Jun 2026 Official leaderboard
42 Kimi K2.7 Code 57.9% Jun 2026 Official leaderboard
43 GPT-5.2 Pro 57.4% xhigh Dec 2025 Official leaderboard
44 GPT-5 56.7% high Aug 2025 Official leaderboard
45 Grok 4.1 Fast 56% Nov 2025 Official leaderboard
46 GLM-5.1 55.1% Apr 2026 Official leaderboard
47 Ox Alpha 55.1% Aug 2026 Official leaderboard
48 Claude Sonnet 4.5 54.3% Sep 2025 Official leaderboard
49 GPT-5.1 53.2% high Nov 2025 Official leaderboard
50 GLM-5 53.2% Feb 2026 Official leaderboard
51 o3 53.1% high Apr 2025 Official leaderboard
52 DeepSeek-V3.2-Speciale 52.6% Dec 2025 Official leaderboard
53 Gemini 2.5 Pro Mar 2025 51.6% Mar 2025 Official leaderboard
54 DeepSeek-V4-Pro 50.9% Apr 2026 Official leaderboard
55 Inkling 50% Jul 2026 Official leaderboard
56 GPT-5.6 Terra 48.9% xhigh Jul 2026 Official leaderboard
57 GLM-4.7 47.7% Dec 2025 Official leaderboard
58 Kimi K2.5 46.8% Jan 2026 Official leaderboard
59 GPT-5.6 Luna 46.8% xhigh Jul 2026 Official leaderboard
60 Claude 3.7 Sonnet 46.4% thinking Feb 2025 Official leaderboard
61 DeepSeek V4 Flash (preview) 46.3% Jun 2026 Official leaderboard
62 GPT-5.2 45.8% high Dec 2025 Official leaderboard
63 MiniMax-M3 45.8% Jun 2026 Official leaderboard
64 Claude Sonnet 4 45.5% thinking May 2025 Official leaderboard
65 o1-preview 41.7% Sep 2024 Official leaderboard
66 Nemotron 3 Ultra 41.7% Jun 2026 Official leaderboard
67 Claude 3.5 Sonnet (October 2024) 41.4% Oct 2024 Official leaderboard
68 Gemini 2.5 Flash Sep 2025 41.2% Sep 2025 Official leaderboard
69 DeepSeek-R1 May 2025 40.8% May 2025 Official leaderboard
70 o1 40.1% high Dec 2024 Official leaderboard
71 DeepSeek-V3.1 40% Aug 2025 Official leaderboard
72 Kimi K2 Thinking 39.6% Nov 2025 Official leaderboard
73 o4-mini 38.7% high Apr 2025 Official leaderboard
74 Grok 3 36.1% Apr 2025 Official leaderboard
75 Qwen 3.6 Flash 35.2% Apr 2026 Official leaderboard
76 MiniMax-M2.1 34.7% Dec 2025 Official leaderboard
77 GPT-4.5 34.5% Feb 2025 Official leaderboard
78 Gemini-exp-1206 31.1% Dec 2024 Official leaderboard
79 Qwen3-235B-A22B 31% Apr 2025 Official leaderboard
80 DeepSeek-R1 30.9% Jan 2025 Official leaderboard
81 Gemini 2.0 Flash Thinking Jan 2025 30.7% Jan 2025 Official leaderboard
82 Llama 4 Maverick 27.7% Apr 2025 Official leaderboard
83 Claude 3.5 Sonnet 27.5% Jun 2024 Official leaderboard
84 DeepSeek-V3 Mar 2025 27.2% Mar 2025 Official leaderboard
85 Gemini 1.5 Pro Sep 2024 27.1% Sep 2024 Official leaderboard
86 GPT-4.1 27% Apr 2025 Official leaderboard
87 Kimi K2 Jul 2025 26.3% Jul 2025 Official leaderboard
88 GPT-4 Turbo Apr 2024 25.1% Apr 2024 Official leaderboard
89 MiniMax-M2 25% Oct 2025 Official leaderboard
90 Claude 3 Opus 23.5% Feb 2024 Official leaderboard
91 Llama 3.1-405B 23% Jul 2024 Official leaderboard
92 o3-mini 22.8% high Jan 2025 Official leaderboard
93 Grok-2 Aug 2024 22.7% Aug 2024 Official leaderboard
94 Mistral Large v2 22.5% Nov 2024 Official leaderboard
95 gpt-oss-120b 22.1% Aug 2025 Official leaderboard
96 Mistral Large 3 20.4% Dec 2025 Official leaderboard
97 Llama 3.3 70B Instruct 19.9% Dec 2024 Official leaderboard
98 Gemini 2.0 Flash Dec 2024 18.9% Dec 2024 Official leaderboard
99 DeepSeek-V3 18.9% Dec 2024 Official leaderboard
100 o1-mini 18.1% Sep 2024 Official leaderboard
101 GPT-4o Aug 2024 17.8% Aug 2024 Official leaderboard
102 Command R+ 17.4% Aug 2024 Official leaderboard
103 GPT-4o mini 10.7% Jul 2024 Official leaderboard
104 GPT-3.5 Turbo Jun 2023 8.00% Jun 2023 Official leaderboard

Epoch AI has scores for 82 models. We plot every source in the charts below.

11 models have held number one since its release in Oct 2024

Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and Epoch AI’s runs.

OpenAI Anthropic Google DeepMind
Own leaderboard Passed 85% in Sep 2026, by Claude Fable 5.1
Claude Opus 5.5 88.4%
Epoch AI runs Not tested since Aug 2026
Claude Fable 5 81.9%

The best score passed 85% in Sep 2026

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Run or collected by Epoch AI Best so far
0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 Near the ceiling (85%) Random guessing Claude Opus 5.5 Claude Fable 5 Gemini 3 Pro Gemini 2.5 Pro o3 Claude 3.7 Sonnet o1-preview 0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 Near the ceiling (85%) Random guessing Claude Opus 5.5 Claude Fable 5

No lab model card reports it yet

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.

More in Science and reasoning

All 88 in Science and reasoning
  • GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics. Gemini 4 Argon 96.4%
  • Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
  • FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics. Intern-S2-397B 87%
  • BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Opus 5 90.1%
  • HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels. Gemini 3 Pro 78.3%
  • LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%