Benchmark Almanac

EKI Propaganda Resistance

Claude Opus 5 97.6 leads since July 2026
Not in model cards

Models answer 75 questions in Estonian, English, and Russian without web search. Claude Opus 4.5 grades how each answer handles misleading premises and propaganda.

The institute publishes scores for 101 models. It combines ratings from 1 to 5 using a geometric mean and rescales the result to 100.

Released
June 2026
Kind
Deception and alignment
Level
Everyday
Scored
Graded by a model
Answers
Open-ended
Measures
Capability
Language
Estonian, English, Russian
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
Strongest model tested
Claude Fable 5.2 #2 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
101
Entries on its own leaderboard, 96 models we track

Claude Opus 5 leads with 97.6 Claude Opus 5 leads with 97.6

Best score per entry from the benchmark's own leaderboard, read on November 20th, 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. This view shows only models we track, each at its best.

101 entries. Scroll sideways for the rest.

  1. 97.6 Claude Opus 5
  2. 97.4 Grok 4.7
  3. 97.0 Claude Fable 5.1
  4. 95.7 Grok 4.6
  5. 95.7 GLM-5.3
  6. 95.2 Grok 4.5
  7. 95.2 Claude Fable 5
  8. 94.9 Claude Opus 4.7
  9. 94.4 GLM-5.3-Flash
  10. 94.3 Claude Fable 5.2
  11. 93.4 Claude Sonnet 5.5
  12. 93.2 Claude Opus 5.5
  13. 93.0 GPT-5.6 Terra
  14. 92.8 GPT-5.6 Sol
  15. 92.8 Kimi K3
  16. 92.7 Claude Opus 4.8
  17. 92.7 Nemotron 3 Super 120B
  18. 92.5 Claude Sonnet 5
  19. 92.2 GPT-6 Astra
  20. 92.1 Qwen 3.6 Plus
  21. 91.2 Claude Sonnet 4.6
  22. 90.7 Muse Spark Open
  23. 90.6 Muse Spark 1.3
  24. 90.2 GPT-6 Luna
  25. 90.1 Mistral Large 4
  26. 90.0 MiMo-V2.6 Pro
  27. 89.9 DeepSeek V4.1 Flash
  28. 89.8 Claude Opus 4.6
  29. 89.5 GPT-5.6 Luna
  30. 89.5 GPT-6.1 Terra
  31. 89.2 Claude Haiku 5.5
  32. 89.0 GPT-6 Sol
  33. 89.0 Claude Opus 4.5
  34. 88.9 GPT-5.4
  35. 88.7 Qwen3.6 27B
  36. 88.2 Claude Sonnet 4.5
  37. 88.2 MiniMax-M3
  38. 87.9 GPT-5.5
  39. 87.5 Grok 4.20
  40. 87.4 Qwen3.8 Flash
  41. 87.3 Qwen3.7-Max
  42. 87.2 Nemotron 3 Ultra
  43. 86.8 Qwen 3.8 Max
  44. 86.5 Step 3.7 Flash
  45. 85.7 Tencent Hy4 preview
  46. 85.7 Qwen3.5-27B
  47. 84.3 GPT-5.4 Mini
  48. 83.7 Claude Haiku 4.5
  49. 82.9 DeepSeek-V3.2
  50. 82.1 MiMo-V2.6 Flash
  51. 82.0 Gemini 2.5 Pro
  52. 81.7 MiMo-V2-Pro
  53. 80.8 Muse Spark 1.2
  54. 79.5 Gemini 4 Argon
  55. 79.2 GPT-5
  56. 79.0 Gemini 3.1 Pro
  57. 78.4 GPT-5.2
  58. 78.1 Step 3.5 Flash
  59. 77.3 Gemini 3.8 Flash
  60. 76.8 Claude 3.7 Sonnet
  61. 76.4 Muse Glimmer
  62. 75.0 Kimi K2
  63. 74.9 Gemini 3.6 Flash
  64. 74.4 Claude Sonnet 4
  65. 74.3 Gemini 3.7 Flash
  66. 74.1 Claude Opus 4
  67. 73.1 Claude 3.5 Haiku
  68. 73.0 Gemini 2.5 Flash
  69. 73.0 Gemini 3.5 Flash
  70. 72.6 DeepSeek-V4-Flash
  71. 72.3 DeepSeek-V4-Pro
  72. 70.4 GLM-5.2
  73. 69.6 MiniMax-M2.7
  74. 69.2 Qwen 3.8 27B
  75. 67.4 Grok 3
  76. 66.5 Gemma 2 27B
  77. 66.3 GPT-4o
  78. 65.2 GLM-4.7
  79. 65.1 Grok 4 Fast
  80. 64.9 DeepSeek-V3
  81. 64.8 Gemini 3 Flash
  82. 64.7 Gemini 2.0 Flash
  83. 64.7 Gemma 3 27B
  84. 64.5 GLM-5
  85. 64.3 Gemma 4 31B IT
  86. 63.7 Laguna S 2.1
  87. 63.6 GPT-4.1
  88. 62.5 GPT-4
  89. 62.3 Mistral Medium 3.5
  90. 62.0 Claude 3 Haiku
  91. 61.8 Mistral Large 2512
  92. 61.3 Grok-3 mini
  93. 59.8 Llama 4 Maverick
  94. 58.4 Mistral Large 2411
  95. 58.3 Command R+ (08-2024)
  96. 56.4 Mistral Small 4
  97. 55.5 Llama 3.3 70B Instruct
  98. 54.5 GPT-4o mini
  99. 52.9 Llama 3.1 70B Instruct
  100. 52.5 gpt-oss-120b
  101. 48.2 GPT-3.5 Turbo
  1. 1 Claude Opus 5 97.6
  2. 2 Grok 4.7 97.4
  3. 3 Claude Fable 5.1 97.0
  4. 4 Grok 4.6 95.7
  5. 5 GLM-5.3 95.7
  6. 6 Grok 4.5 95.2
  7. 7 Claude Fable 5 95.2
  8. 8 Claude Opus 4.7 94.9
  9. 9 GLM-5.3-Flash 94.4
  10. 10 Claude Fable 5.2 94.3
  11. 11 Claude Sonnet 5.5 93.4
  12. 12 Claude Opus 5.5 93.2
  13. 13 GPT-5.6 Terra 93.0
  14. 14 GPT-5.6 Sol 92.8
  15. 15 Kimi K3 92.8
  16. 16 Claude Opus 4.8 92.7
  17. 17 Nemotron 3 Super 120B 92.7
  18. 18 Claude Sonnet 5 92.5
  19. 19 GPT-6 Astra 92.2
  20. 20 Qwen 3.6 Plus 92.1
  21. 21 Claude Sonnet 4.6 91.2
  22. 22 Muse Spark Open 90.7
  23. 23 Muse Spark 1.3 90.6
  24. 24 GPT-6 Luna 90.2
  25. 25 Mistral Large 4 90.1
  26. 26 MiMo-V2.6 Pro 90.0
  27. 27 DeepSeek V4.1 Flash 89.9
  28. 28 Claude Opus 4.6 89.8
  29. 29 GPT-5.6 Luna 89.5
  30. 30 GPT-6.1 Terra 89.5
  31. 31 Claude Haiku 5.5 89.2
  32. 32 GPT-6 Sol 89.0
  33. 33 Claude Opus 4.5 89.0
  34. 34 GPT-5.4 88.9
  35. 35 Qwen3.6 27B 88.7
  36. 36 Claude Sonnet 4.5 88.2
  37. 37 MiniMax-M3 88.2
  38. 38 GPT-5.5 87.9
  39. 39 Grok 4.20 87.5
  40. 40 Qwen3.8 Flash 87.4
  41. 41 Qwen3.7-Max 87.3
  42. 42 Nemotron 3 Ultra 87.2
  43. 43 Qwen 3.8 Max 86.8
  44. 44 Step 3.7 Flash 86.5
  45. 45 Tencent Hy4 preview 85.7
  46. 46 Qwen3.5-27B 85.7
  47. 47 GPT-5.4 Mini 84.3
  48. 48 Claude Haiku 4.5 83.7
  49. 49 DeepSeek-V3.2 82.9
  50. 50 MiMo-V2.6 Flash 82.1
  51. 51 Gemini 2.5 Pro Jun 2025 82.0
  52. 52 MiMo-V2-Pro 81.7
  53. 53 Muse Spark 1.2 80.8
  54. 54 Gemini 4 Argon 79.5
  55. 55 GPT-5 79.2
  56. 56 Gemini 3.1 Pro 79.0
  57. 57 GPT-5.2 78.4
  58. 58 Step 3.5 Flash 78.1
  59. 59 Gemini 3.8 Flash 77.3
  60. 60 Claude 3.7 Sonnet 76.8
  61. 61 Muse Glimmer 76.4
  62. 62 Kimi K2 Sep 2025 75.0
  63. 63 Gemini 3.6 Flash 74.9
  64. 64 Claude Sonnet 4 74.4
  65. 65 Gemini 3.7 Flash 74.3
  66. 66 Claude Opus 4 74.1
  67. 67 Claude 3.5 Haiku 73.1
  68. 68 Gemini 2.5 Flash Sep 2025 73.0
  69. 69 Gemini 3.5 Flash 73.0
  70. 70 DeepSeek-V4-Flash 72.6
  71. 71 DeepSeek-V4-Pro 72.3
  72. 72 GLM-5.2 70.4
  73. 73 MiniMax-M2.7 69.6
  74. 74 Qwen 3.8 27B 69.2
  75. 75 Grok 3 67.4
  76. 76 Gemma 2 27B 66.5
  77. 77 GPT-4o Mar 2025 66.3
  78. 78 GLM-4.7 65.2
  79. 79 Grok 4 Fast 65.1
  80. 80 DeepSeek-V3 Mar 2025 64.9
  81. 81 Gemini 3 Flash 64.8
  82. 82 Gemini 2.0 Flash Feb 2025 64.7
  83. 83 Gemma 3 27B 64.7
  84. 84 GLM-5 64.5
  85. 85 Gemma 4 31B IT 64.3
  86. 86 Laguna S 2.1 63.7
  87. 87 GPT-4.1 63.6
  88. 88 GPT-4 Jun 2023 62.5
  89. 89 Mistral Medium 3.5 62.3
  90. 90 Claude 3 Haiku 62.0
  91. 91 Mistral Large 2512 61.8
  92. 92 Grok-3 mini 61.3
  93. 93 Llama 4 Maverick 59.8
  94. 94 Mistral Large 2411 58.4
  95. 95 Command R+ (08-2024) 58.3
  96. 96 Mistral Small 4 56.4
  97. 97 Llama 3.3 70B Instruct 55.5
  98. 98 GPT-4o mini 54.5
  99. 99 Llama 3.1 70B Instruct 52.9
  100. 100 gpt-oss-120b 52.5
  101. 101 GPT-3.5 Turbo Jan 2024 48.2

Source: Official EKI Propaganda Resistance leaderboard

96 models. Scroll sideways for the rest.

  1. 97.6 Claude Opus 5
  2. 97.4 Grok 4.7
  3. 97.0 Claude Fable 5.1
  4. 95.7 Grok 4.6
  5. 95.7 GLM-5.3
  6. 95.2 Grok 4.5
  7. 95.2 Claude Fable 5
  8. 94.9 Claude Opus 4.7
  9. 94.4 GLM-5.3-Flash
  10. 94.3 Claude Fable 5.2
  11. 93.4 Claude Sonnet 5.5
  12. 93.2 Claude Opus 5.5
  13. 93.0 GPT-5.6 Terra
  14. 92.8 GPT-5.6 Sol
  15. 92.8 Kimi K3
  16. 92.7 Claude Opus 4.8
  17. 92.5 Claude Sonnet 5
  18. 92.2 GPT-6 Astra
  19. 92.1 Qwen 3.6 Plus
  20. 91.2 Claude Sonnet 4.6
  21. 90.7 Muse Spark Open
  22. 90.6 Muse Spark 1.3
  23. 90.2 GPT-6 Luna
  24. 90.1 Mistral Large 4
  25. 90.0 MiMo-V2.6 Pro
  26. 89.9 DeepSeek V4.1 Flash
  27. 89.8 Claude Opus 4.6
  28. 89.5 GPT-5.6 Luna
  29. 89.5 GPT-6.1 Terra
  30. 89.2 Claude Haiku 5.5
  31. 89.0 GPT-6 Sol
  32. 89.0 Claude Opus 4.5
  33. 88.9 GPT-5.4
  34. 88.7 Qwen3.6 27B
  35. 88.2 Claude Sonnet 4.5
  36. 88.2 MiniMax-M3
  37. 87.9 GPT-5.5
  38. 87.5 Grok 4.20
  39. 87.4 Qwen3.8 Flash
  40. 87.3 Qwen3.7-Max
  41. 87.2 Nemotron 3 Ultra
  42. 86.8 Qwen 3.8 Max
  43. 86.5 Step 3.7 Flash
  44. 85.7 Tencent Hy4 preview
  45. 85.7 Qwen3.5-27B
  46. 84.3 GPT-5.4 Mini
  47. 83.7 Claude Haiku 4.5
  48. 82.9 DeepSeek-V3.2
  49. 82.1 MiMo-V2.6 Flash
  50. 82.0 Gemini 2.5 Pro
  51. 81.7 MiMo-V2-Pro
  52. 80.8 Muse Spark 1.2
  53. 79.5 Gemini 4 Argon
  54. 79.2 GPT-5
  55. 79.0 Gemini 3.1 Pro
  56. 78.4 GPT-5.2
  57. 78.1 Step 3.5 Flash
  58. 77.3 Gemini 3.8 Flash
  59. 76.8 Claude 3.7 Sonnet
  60. 76.4 Muse Glimmer
  61. 75.0 Kimi K2
  62. 74.9 Gemini 3.6 Flash
  63. 74.4 Claude Sonnet 4
  64. 74.3 Gemini 3.7 Flash
  65. 74.1 Claude Opus 4
  66. 73.1 Claude 3.5 Haiku
  67. 73.0 Gemini 2.5 Flash
  68. 73.0 Gemini 3.5 Flash
  69. 72.6 DeepSeek-V4-Flash
  70. 72.3 DeepSeek-V4-Pro
  71. 70.4 GLM-5.2
  72. 69.6 MiniMax-M2.7
  73. 69.2 Qwen 3.8 27B
  74. 67.4 Grok 3
  75. 66.5 Gemma 2 27B
  76. 66.3 GPT-4o
  77. 65.2 GLM-4.7
  78. 65.1 Grok 4 Fast
  79. 64.9 DeepSeek-V3
  80. 64.8 Gemini 3 Flash
  81. 64.7 Gemini 2.0 Flash
  82. 64.7 Gemma 3 27B
  83. 64.5 GLM-5
  84. 64.3 Gemma 4 31B IT
  85. 63.7 Laguna S 2.1
  86. 63.6 GPT-4.1
  87. 62.5 GPT-4
  88. 62.3 Mistral Medium 3.5
  89. 62.0 Claude 3 Haiku
  90. 61.3 Grok-3 mini
  91. 59.8 Llama 4 Maverick
  92. 56.4 Mistral Small 4
  93. 55.5 Llama 3.3 70B Instruct
  94. 54.5 GPT-4o mini
  95. 52.5 gpt-oss-120b
  96. 48.2 GPT-3.5 Turbo
  1. 1 Claude Opus 5 97.6
  2. 2 Grok 4.7 97.4
  3. 3 Claude Fable 5.1 97.0
  4. 4 Grok 4.6 95.7
  5. 5 GLM-5.3 95.7
  6. 6 Grok 4.5 95.2
  7. 7 Claude Fable 5 95.2
  8. 8 Claude Opus 4.7 94.9
  9. 9 GLM-5.3-Flash 94.4
  10. 10 Claude Fable 5.2 94.3
  11. 11 Claude Sonnet 5.5 93.4
  12. 12 Claude Opus 5.5 93.2
  13. 13 GPT-5.6 Terra 93.0
  14. 14 GPT-5.6 Sol 92.8
  15. 15 Kimi K3 92.8
  16. 16 Claude Opus 4.8 92.7
  17. 17 Claude Sonnet 5 92.5
  18. 18 GPT-6 Astra 92.2
  19. 19 Qwen 3.6 Plus 92.1
  20. 20 Claude Sonnet 4.6 91.2
  21. 21 Muse Spark Open 90.7
  22. 22 Muse Spark 1.3 90.6
  23. 23 GPT-6 Luna 90.2
  24. 24 Mistral Large 4 90.1
  25. 25 MiMo-V2.6 Pro 90.0
  26. 26 DeepSeek V4.1 Flash 89.9
  27. 27 Claude Opus 4.6 89.8
  28. 28 GPT-5.6 Luna 89.5
  29. 29 GPT-6.1 Terra 89.5
  30. 30 Claude Haiku 5.5 89.2
  31. 31 GPT-6 Sol 89.0
  32. 32 Claude Opus 4.5 89.0
  33. 33 GPT-5.4 88.9
  34. 34 Qwen3.6 27B 88.7
  35. 35 Claude Sonnet 4.5 88.2
  36. 36 MiniMax-M3 88.2
  37. 37 GPT-5.5 87.9
  38. 38 Grok 4.20 87.5
  39. 39 Qwen3.8 Flash 87.4
  40. 40 Qwen3.7-Max 87.3
  41. 41 Nemotron 3 Ultra 87.2
  42. 42 Qwen 3.8 Max 86.8
  43. 43 Step 3.7 Flash 86.5
  44. 44 Tencent Hy4 preview 85.7
  45. 45 Qwen3.5-27B 85.7
  46. 46 GPT-5.4 Mini 84.3
  47. 47 Claude Haiku 4.5 83.7
  48. 48 DeepSeek-V3.2 82.9
  49. 49 MiMo-V2.6 Flash 82.1
  50. 50 Gemini 2.5 Pro Jun 2025 82.0
  51. 51 MiMo-V2-Pro 81.7
  52. 52 Muse Spark 1.2 80.8
  53. 53 Gemini 4 Argon 79.5
  54. 54 GPT-5 79.2
  55. 55 Gemini 3.1 Pro 79.0
  56. 56 GPT-5.2 78.4
  57. 57 Step 3.5 Flash 78.1
  58. 58 Gemini 3.8 Flash 77.3
  59. 59 Claude 3.7 Sonnet 76.8
  60. 60 Muse Glimmer 76.4
  61. 61 Kimi K2 Sep 2025 75.0
  62. 62 Gemini 3.6 Flash 74.9
  63. 63 Claude Sonnet 4 74.4
  64. 64 Gemini 3.7 Flash 74.3
  65. 65 Claude Opus 4 74.1
  66. 66 Claude 3.5 Haiku 73.1
  67. 67 Gemini 2.5 Flash Sep 2025 73.0
  68. 68 Gemini 3.5 Flash 73.0
  69. 69 DeepSeek-V4-Flash 72.6
  70. 70 DeepSeek-V4-Pro 72.3
  71. 71 GLM-5.2 70.4
  72. 72 MiniMax-M2.7 69.6
  73. 73 Qwen 3.8 27B 69.2
  74. 74 Grok 3 67.4
  75. 75 Gemma 2 27B 66.5
  76. 76 GPT-4o Mar 2025 66.3
  77. 77 GLM-4.7 65.2
  78. 78 Grok 4 Fast 65.1
  79. 79 DeepSeek-V3 Mar 2025 64.9
  80. 80 Gemini 3 Flash 64.8
  81. 81 Gemini 2.0 Flash Feb 2025 64.7
  82. 82 Gemma 3 27B 64.7
  83. 83 GLM-5 64.5
  84. 84 Gemma 4 31B IT 64.3
  85. 85 Laguna S 2.1 63.7
  86. 86 GPT-4.1 63.6
  87. 87 GPT-4 Jun 2023 62.5
  88. 88 Mistral Medium 3.5 62.3
  89. 89 Claude 3 Haiku 62.0
  90. 90 Grok-3 mini 61.3
  91. 91 Llama 4 Maverick 59.8
  92. 92 Mistral Small 4 56.4
  93. 93 Llama 3.3 70B Instruct 55.5
  94. 94 GPT-4o mini 54.5
  95. 95 gpt-oss-120b 52.5
  96. 96 GPT-3.5 Turbo Jan 2024 48.2

Source: Official EKI Propaganda Resistance leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 5 97.6 Jul 2026 Official leaderboard
2 Grok 4.7 97.4 Sep 2026 Official leaderboard
3 Claude Fable 5.1 97.0 Sep 2026 Official leaderboard
4 Grok 4.6 95.7 Aug 2026 Official leaderboard
5 GLM-5.3 95.7 Aug 2026 Official leaderboard
6 Grok 4.5 95.2 Jul 2026 Official leaderboard
7 Claude Fable 5 95.2 Jun 2026 Official leaderboard
8 Claude Opus 4.7 94.9 Apr 2026 Official leaderboard
9 GLM-5.3-Flash 94.4 Aug 2026 Official leaderboard
10 Claude Fable 5.2 94.3 Nov 2026 Official leaderboard
11 Claude Sonnet 5.5 93.4 Sep 2026 Official leaderboard
12 Claude Opus 5.5 93.2 Sep 2026 Official leaderboard
13 GPT-5.6 Terra 93.0 Jul 2026 Official leaderboard
14 GPT-5.6 Sol 92.8 Jul 2026 Official leaderboard
15 Kimi K3 92.8 Jul 2026 Official leaderboard
16 Claude Opus 4.8 92.7 May 2026 Official leaderboard
17 Nemotron 3 Super 120B open weights 92.7 Official leaderboard
18 Claude Sonnet 5 92.5 Jun 2026 Official leaderboard
19 GPT-6 Astra 92.2 Sep 2026 Official leaderboard
20 Qwen 3.6 Plus 92.1 Mar 2026 Official leaderboard
21 Claude Sonnet 4.6 91.2 Feb 2026 Official leaderboard
22 Muse Spark Open 90.7 Oct 2026 Official leaderboard
23 Muse Spark 1.3 90.6 Sep 2026 Official leaderboard
24 GPT-6 Luna 90.2 Sep 2026 Official leaderboard
25 Mistral Large 4 90.1 Oct 2026 Official leaderboard
26 MiMo-V2.6 Pro 90.0 Sep 2026 Official leaderboard
27 DeepSeek V4.1 Flash 89.9 Sep 2026 Official leaderboard
28 Claude Opus 4.6 89.8 Feb 2026 Official leaderboard
29 GPT-5.6 Luna 89.5 Jul 2026 Official leaderboard
30 GPT-6.1 Terra 89.5 Oct 2026 Official leaderboard
31 Claude Haiku 5.5 89.2 Oct 2026 Official leaderboard
32 GPT-6 Sol 89.0 Sep 2026 Official leaderboard
33 Claude Opus 4.5 89.0 Nov 2025 Official leaderboard
34 GPT-5.4 88.9 Mar 2026 Official leaderboard
35 Qwen3.6 27B 88.7 Apr 2026 Official leaderboard
36 Claude Sonnet 4.5 88.2 Sep 2025 Official leaderboard
37 MiniMax-M3 88.2 Jun 2026 Official leaderboard
38 GPT-5.5 87.9 Apr 2026 Official leaderboard
39 Grok 4.20 87.5 Feb 2026 Official leaderboard
40 Qwen3.8 Flash 87.4 Aug 2026 Official leaderboard
41 Qwen3.7-Max 87.3 May 2026 Official leaderboard
42 Nemotron 3 Ultra 87.2 Jun 2026 Official leaderboard
43 Qwen 3.8 Max 86.8 Aug 2026 Official leaderboard
44 Step 3.7 Flash 86.5 May 2026 Official leaderboard
45 Tencent Hy4 preview 85.7 Aug 2026 Official leaderboard
46 Qwen3.5-27B 85.7 Feb 2026 Official leaderboard
47 GPT-5.4 Mini 84.3 Mar 2026 Official leaderboard
48 Claude Haiku 4.5 83.7 Oct 2025 Official leaderboard
49 DeepSeek-V3.2 82.9 Dec 2025 Official leaderboard
50 MiMo-V2.6 Flash 82.1 Sep 2026 Official leaderboard
51 Gemini 2.5 Pro Jun 2025 82.0 Jun 2025 Official leaderboard
52 MiMo-V2-Pro 81.7 Mar 2026 Official leaderboard
53 Muse Spark 1.2 80.8 Aug 2026 Official leaderboard
54 Gemini 4 Argon 79.5 Sep 2026 Official leaderboard
55 GPT-5 79.2 Aug 2025 Official leaderboard
56 Gemini 3.1 Pro 79.0 Feb 2026 Official leaderboard
57 GPT-5.2 78.4 Dec 2025 Official leaderboard
58 Step 3.5 Flash 78.1 Feb 2026 Official leaderboard
59 Gemini 3.8 Flash 77.3 Sep 2026 Official leaderboard
60 Claude 3.7 Sonnet 76.8 Feb 2025 Official leaderboard
61 Muse Glimmer 76.4 Aug 2026 Official leaderboard
62 Kimi K2 Sep 2025 75.0 Sep 2025 Official leaderboard
63 Gemini 3.6 Flash 74.9 Jul 2026 Official leaderboard
64 Claude Sonnet 4 74.4 May 2025 Official leaderboard
65 Gemini 3.7 Flash 74.3 Aug 2026 Official leaderboard
66 Claude Opus 4 74.1 May 2025 Official leaderboard
67 Claude 3.5 Haiku 73.1 Oct 2024 Official leaderboard
68 Gemini 2.5 Flash Sep 2025 73.0 Sep 2025 Official leaderboard
69 Gemini 3.5 Flash 73.0 May 2026 Official leaderboard
70 DeepSeek-V4-Flash 72.6 Apr 2026 Official leaderboard
71 DeepSeek-V4-Pro 72.3 Apr 2026 Official leaderboard
72 GLM-5.2 70.4 Jun 2026 Official leaderboard
73 MiniMax-M2.7 69.6 Mar 2026 Official leaderboard
74 Qwen 3.8 27B 69.2 Aug 2026 Official leaderboard
75 Grok 3 67.4 Apr 2025 Official leaderboard
76 Gemma 2 27B 66.5 Jun 2024 Official leaderboard
77 GPT-4o Mar 2025 66.3 Mar 2025 Official leaderboard
78 GLM-4.7 65.2 Dec 2025 Official leaderboard
79 Grok 4 Fast 65.1 Sep 2025 Official leaderboard
80 DeepSeek-V3 Mar 2025 64.9 Mar 2025 Official leaderboard
81 Gemini 3 Flash 64.8 Dec 2025 Official leaderboard
82 Gemini 2.0 Flash Feb 2025 64.7 Feb 2025 Official leaderboard
83 Gemma 3 27B 64.7 Mar 2025 Official leaderboard
84 GLM-5 64.5 Feb 2026 Official leaderboard
85 Gemma 4 31B IT 64.3 Apr 2026 Official leaderboard
86 Laguna S 2.1 63.7 Jul 2026 Official leaderboard
87 GPT-4.1 63.6 Apr 2025 Official leaderboard
88 GPT-4 Jun 2023 62.5 Jun 2023 Official leaderboard
89 Mistral Medium 3.5 62.3 Apr 2026 Official leaderboard
90 Claude 3 Haiku 62.0 Mar 2024 Official leaderboard
91 Mistral Large 2512 open weights 61.8 Official leaderboard
92 Grok-3 mini 61.3 Apr 2025 Official leaderboard
93 Llama 4 Maverick 59.8 Apr 2025 Official leaderboard
94 Mistral Large 2411 open weights 58.4 Official leaderboard
95 Command R+ (08-2024) 58.3 Official leaderboard
96 Mistral Small 4 56.4 Mar 2026 Official leaderboard
97 Llama 3.3 70B Instruct 55.5 Dec 2024 Official leaderboard
98 GPT-4o mini 54.5 Jul 2024 Official leaderboard
99 Llama 3.1 70B Instruct open weights 52.9 Official leaderboard
100 gpt-oss-120b 52.5 Aug 2025 Official leaderboard
101 GPT-3.5 Turbo Jan 2024 48.2 Jan 2024 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 5 97.6 Jul 2026 Official leaderboard
2 Grok 4.7 97.4 Sep 2026 Official leaderboard
3 Claude Fable 5.1 97.0 Sep 2026 Official leaderboard
4 Grok 4.6 95.7 Aug 2026 Official leaderboard
5 GLM-5.3 95.7 Aug 2026 Official leaderboard
6 Grok 4.5 95.2 Jul 2026 Official leaderboard
7 Claude Fable 5 95.2 Jun 2026 Official leaderboard
8 Claude Opus 4.7 94.9 Apr 2026 Official leaderboard
9 GLM-5.3-Flash 94.4 Aug 2026 Official leaderboard
10 Claude Fable 5.2 94.3 Nov 2026 Official leaderboard
11 Claude Sonnet 5.5 93.4 Sep 2026 Official leaderboard
12 Claude Opus 5.5 93.2 Sep 2026 Official leaderboard
13 GPT-5.6 Terra 93.0 Jul 2026 Official leaderboard
14 GPT-5.6 Sol 92.8 Jul 2026 Official leaderboard
15 Kimi K3 92.8 Jul 2026 Official leaderboard
16 Claude Opus 4.8 92.7 May 2026 Official leaderboard
17 Claude Sonnet 5 92.5 Jun 2026 Official leaderboard
18 GPT-6 Astra 92.2 Sep 2026 Official leaderboard
19 Qwen 3.6 Plus 92.1 Mar 2026 Official leaderboard
20 Claude Sonnet 4.6 91.2 Feb 2026 Official leaderboard
21 Muse Spark Open 90.7 Oct 2026 Official leaderboard
22 Muse Spark 1.3 90.6 Sep 2026 Official leaderboard
23 GPT-6 Luna 90.2 Sep 2026 Official leaderboard
24 Mistral Large 4 90.1 Oct 2026 Official leaderboard
25 MiMo-V2.6 Pro 90.0 Sep 2026 Official leaderboard
26 DeepSeek V4.1 Flash 89.9 Sep 2026 Official leaderboard
27 Claude Opus 4.6 89.8 Feb 2026 Official leaderboard
28 GPT-5.6 Luna 89.5 Jul 2026 Official leaderboard
29 GPT-6.1 Terra 89.5 Oct 2026 Official leaderboard
30 Claude Haiku 5.5 89.2 Oct 2026 Official leaderboard
31 GPT-6 Sol 89.0 Sep 2026 Official leaderboard
32 Claude Opus 4.5 89.0 Nov 2025 Official leaderboard
33 GPT-5.4 88.9 Mar 2026 Official leaderboard
34 Qwen3.6 27B 88.7 Apr 2026 Official leaderboard
35 Claude Sonnet 4.5 88.2 Sep 2025 Official leaderboard
36 MiniMax-M3 88.2 Jun 2026 Official leaderboard
37 GPT-5.5 87.9 Apr 2026 Official leaderboard
38 Grok 4.20 87.5 Feb 2026 Official leaderboard
39 Qwen3.8 Flash 87.4 Aug 2026 Official leaderboard
40 Qwen3.7-Max 87.3 May 2026 Official leaderboard
41 Nemotron 3 Ultra 87.2 Jun 2026 Official leaderboard
42 Qwen 3.8 Max 86.8 Aug 2026 Official leaderboard
43 Step 3.7 Flash 86.5 May 2026 Official leaderboard
44 Tencent Hy4 preview 85.7 Aug 2026 Official leaderboard
45 Qwen3.5-27B 85.7 Feb 2026 Official leaderboard
46 GPT-5.4 Mini 84.3 Mar 2026 Official leaderboard
47 Claude Haiku 4.5 83.7 Oct 2025 Official leaderboard
48 DeepSeek-V3.2 82.9 Dec 2025 Official leaderboard
49 MiMo-V2.6 Flash 82.1 Sep 2026 Official leaderboard
50 Gemini 2.5 Pro Jun 2025 82.0 Jun 2025 Official leaderboard
51 MiMo-V2-Pro 81.7 Mar 2026 Official leaderboard
52 Muse Spark 1.2 80.8 Aug 2026 Official leaderboard
53 Gemini 4 Argon 79.5 Sep 2026 Official leaderboard
54 GPT-5 79.2 Aug 2025 Official leaderboard
55 Gemini 3.1 Pro 79.0 Feb 2026 Official leaderboard
56 GPT-5.2 78.4 Dec 2025 Official leaderboard
57 Step 3.5 Flash 78.1 Feb 2026 Official leaderboard
58 Gemini 3.8 Flash 77.3 Sep 2026 Official leaderboard
59 Claude 3.7 Sonnet 76.8 Feb 2025 Official leaderboard
60 Muse Glimmer 76.4 Aug 2026 Official leaderboard
61 Kimi K2 Sep 2025 75.0 Sep 2025 Official leaderboard
62 Gemini 3.6 Flash 74.9 Jul 2026 Official leaderboard
63 Claude Sonnet 4 74.4 May 2025 Official leaderboard
64 Gemini 3.7 Flash 74.3 Aug 2026 Official leaderboard
65 Claude Opus 4 74.1 May 2025 Official leaderboard
66 Claude 3.5 Haiku 73.1 Oct 2024 Official leaderboard
67 Gemini 2.5 Flash Sep 2025 73.0 Sep 2025 Official leaderboard
68 Gemini 3.5 Flash 73.0 May 2026 Official leaderboard
69 DeepSeek-V4-Flash 72.6 Apr 2026 Official leaderboard
70 DeepSeek-V4-Pro 72.3 Apr 2026 Official leaderboard
71 GLM-5.2 70.4 Jun 2026 Official leaderboard
72 MiniMax-M2.7 69.6 Mar 2026 Official leaderboard
73 Qwen 3.8 27B 69.2 Aug 2026 Official leaderboard
74 Grok 3 67.4 Apr 2025 Official leaderboard
75 Gemma 2 27B 66.5 Jun 2024 Official leaderboard
76 GPT-4o Mar 2025 66.3 Mar 2025 Official leaderboard
77 GLM-4.7 65.2 Dec 2025 Official leaderboard
78 Grok 4 Fast 65.1 Sep 2025 Official leaderboard
79 DeepSeek-V3 Mar 2025 64.9 Mar 2025 Official leaderboard
80 Gemini 3 Flash 64.8 Dec 2025 Official leaderboard
81 Gemini 2.0 Flash Feb 2025 64.7 Feb 2025 Official leaderboard
82 Gemma 3 27B 64.7 Mar 2025 Official leaderboard
83 GLM-5 64.5 Feb 2026 Official leaderboard
84 Gemma 4 31B IT 64.3 Apr 2026 Official leaderboard
85 Laguna S 2.1 63.7 Jul 2026 Official leaderboard
86 GPT-4.1 63.6 Apr 2025 Official leaderboard
87 GPT-4 Jun 2023 62.5 Jun 2023 Official leaderboard
88 Mistral Medium 3.5 62.3 Apr 2026 Official leaderboard
89 Claude 3 Haiku 62.0 Mar 2024 Official leaderboard
90 Grok-3 mini 61.3 Apr 2025 Official leaderboard
91 Llama 4 Maverick 59.8 Apr 2025 Official leaderboard
92 Mistral Small 4 56.4 Mar 2026 Official leaderboard
93 Llama 3.3 70B Instruct 55.5 Dec 2024 Official leaderboard
94 GPT-4o mini 54.5 Jul 2024 Official leaderboard
95 gpt-oss-120b 52.5 Aug 2025 Official leaderboard
96 GPT-3.5 Turbo Jan 2024 48.2 Jan 2024 Official leaderboard

4 models have held number one on the official leaderboard since its release in June 2026

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic xAI
Own leaderboard 4 leaders

Source: Official EKI Propaganda Resistance leaderboard

Scores over time

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

Plot with other benchmarks
On the benchmark’s own leaderboard
40.0 60.0 80.0 100.0 2023 2024 2025 2026 2027 40.0 60.0 80.0 100.0 2023 2024 2025 2026 2027

Source: Official EKI Propaganda Resistance leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Safety and alignment

All 636 in Safety and alignment
  • FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
  • StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is. Inkling 98.6%
  • Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. Gemini 4 Carbon 0%
  • Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior. Inkling 2 Mini 8.30%
  • Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions. Inkling 2 Small 4.70%
  • Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior. Inkling 2 Mini 9.80%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (101 entries).
  • Pass Models scored 96 models we track, plus 5 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Claude Fable 5.2.
  • Pass Frontier models 7 of the top 10 have a score, led by Claude Fable 5.2 (#2).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on xn--mdupuu-pxaa.eki.ee: 96 rows for models we track and 5 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check