Benchmark Almanac

Molecular Biology Capabilities Test

Claude Opus 5.5 68.7% leads since September 2026
Claude Opus 5.5 68.7% leads since September 2026
Only in system cards

Models answer expert-level molecular biology questions from SecureBio. Scored as accuracy.

200 text-only questions with a private holdout on everyday molecular biology methods. 14 PhD biologists scored about 34%. Modeled after VCT.

Made by
SecureBio
Kind
Genomics and life sciences
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Strongest model tested
GPT-6 Astra #3 on the capability index
Last in a system card
October 2026
Last model added
October 2026
Scores refreshed
November 12th, 2026
106
Entries on its own leaderboard, 90 models we track
9
System cards report it, from 3 labs

Claude Opus 5.5 leads with 68.7%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 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.

Claude Opus 5.5 leads with 68.7%, claims included

Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.

106 entries. Scroll sideways for the rest.

  1. 68.7 Claude Opus 5.5
  2. 63.9 GPT-6 Astra
  3. 63.2 Gemini 4 Argon
  4. 61.1 GPT-5.6 Sol
  5. 61.1 GPT-6 Sol
  6. 60.7 Kimi K3.1 Pro
  7. 60.2 Claude Opus 5
  8. 59.8 Kimi K2.6
  9. 59.2 Claude Sonnet 5
  10. 59.1 Claude Opus 4
  11. 58.8 Claude Sonnet 5.5
  12. 58.4 Kimi K3
  13. 57.8 Qwen3.8-2.4T-A95B
  14. 57.1 GPT-5.5
  15. 57.0 GPT-5.5 pro Codex
  16. 56.9 GPT-5.6 Terra
  17. 56.7 Claude Opus 4.6
  18. 56.4 Claude Opus 4.8
  19. 56.0 Claude Opus 4.7
  20. 55.9 GPT-5.5 Rosalind
  21. 55.6 Gemini 3 Flash (Dec '25 Preview)
  22. 55.5 GLM-5.2
  23. 55.1 Claude Opus 4.1
  24. 55.0 Gemini 3.7 Flash
  25. 54.8 GPT-5.2
  26. 54.3 Gemini 3.6 Flash
  27. 54.1 Claude Sonnet 4.5
  28. 53.9 DeepSeek V4.1 Flash 1017
  29. 53.0 DeepSeek V4 Pro 0813
  30. 53.0 Qwen 3.8 27B
  31. 52.9 Claude Opus 4.5
  32. 52.5 Qwen3.5 397B-A17B
  33. 52.5 Nemotron 3.5 Super VL
  34. 52.2 Gemini 2.5 Pro
  35. 51.8 Qwen3.7-Max
  36. 51.7 Grok 4
  37. 51.5 GPT-5 mini
  38. 51.3 Mistral Large 4
  39. 51.2 Inkling
  40. 51.1 o1
  41. 50.8 GPT-5
  42. 50.4 Grok 4.6
  43. 49.8 GLM-5.3
  44. 49.6 GPT-5 Pro
  45. 49.2 o3
  46. 49.2 GPT-5.1
  47. 49.2 Beam 1.0
  48. 48.9 o4-mini
  49. 48.9 Gemini 3.8 Flash
  50. 48.7 Gemini 3 Pro (Nov '25 Preview)
  51. 48.3 Gemma 4 31B IT
  52. 48.2 Inkling-Small
  53. 48.1 GPT-5.4
  54. 48.0 GLM-5
  55. 47.5 Gemini 3.1 Pro (Feb 26' Preview)
  56. 47.5 Qwen3.6 27B
  57. 47.2 Qwen 3.6 Plus
  58. 46.8 Claude Sonnet 3.7
  59. 46.5 GPT-4.5
  60. 46.4 Claude Sonnet 4
  61. 46.4 gpt-oss-120b
  62. 46.1 Kimi K2 Thinking
  63. 46.1 Grok 4.7
  64. 45.7 o3-mini
  65. 45.6 Mistral Medium 3.5
  66. 45.1 Grok 4.1 Fast
  67. 45.0 Gemini 3.5 Flash
  68. 44.8 Nemotron 3 Ultra 550B A55B
  69. 44.4 Qwen3-235B-A22B-Thinking
  70. 44.3 Sonar Deep Research
  71. 44.3 Grok 4 Fast
  72. 43.9 Nemotron 3 Super 120B A12B
  73. 43.1 DeepSeek-R1
  74. 42.9 Claude 3.5 Sonnet (October 2024)
  75. 42.5 GPT-4.1
  76. 42.5 GPT-4.1 mini
  77. 42.5 Claude Haiku 4.5
  78. 42.4 Kimi K2.5
  79. 42.2 Qwen3-Next-80B-A3B
  80. 41.4 Qwen3-235B-A22B
  81. 41.0 Grok 3
  82. 41.0 GPT-5 nano
  83. 40.7 Kimi K2 (Jul '25)
  84. 40.4 Kimi K2 (Sep '25)
  85. 40.2 Llama 4 Maverick
  86. 39.9 GPT-4o
  87. 39.8 Gemini 2.5 Flash (Apr '25 Preview)
  88. 39.8 Gemini 2.5 Flash
  89. 39.6 GPT-4o
  90. 39.1 DeepSeek-V3.1
  91. 39.1 Qwen3-Next-80B-A3B-Instruct
  92. 39.0 DeepSeek-V3.2
  93. 38.5 Nemotron 3 Nano 30B-A3B
  94. 37.5 Grok-3 mini
  95. 35.1 Gemini 2.5 Pro (May '25 Preview)
  96. 33.6 DeepSeek-V4-Pro
  97. 33.4 GPT-4 Turbo
  98. 30.8 Magistral Medium 1.0
  99. 28.0 Claude Haiku 3.5
  100. 27.5 GPT-4
  101. 26.1 Nemotron 3.5 Lightning 30B-A3B
  102. 24.5 GPT-4.1 nano
  103. 24.4 Magistral Small
  104. 22.4 GPT-4o mini
  105. 15.5 Muse Glimmer 30B
  106. 11.5 GPT-3.5 Turbo
  1. 1 Claude Opus 5.5 max 68.7%
  2. 2 GPT-6 Astra max 63.9%
  3. 3 Gemini 4 Argon high 63.2%
  4. 4 GPT-5.6 Sol max 61.1%
  5. 5 GPT-6 Sol max 61.1%
  6. 6 Kimi K3.1 Pro max 60.7%
  7. 7 Claude Opus 5 max 60.2%
  8. 8 Kimi K2.6 59.8%
  9. 9 Claude Sonnet 5 max 59.2%
  10. 10 Claude Opus 4 16k 59.1%
  11. 11 Claude Sonnet 5.5 max 58.8%
  12. 12 Kimi K3 max 58.4%
  13. 13 Qwen3.8-2.4T-A95B xhigh 57.8%
  14. 14 GPT-5.5 xhigh 57.1%
  15. 15 GPT-5.5 pro Codex xhigh 57%
  16. 16 GPT-5.6 Terra max 56.9%
  17. 17 Claude Opus 4.6 max 56.7%
  18. 18 Claude Opus 4.8 max 56.4%
  19. 19 Claude Opus 4.7 max 56%
  20. 20 GPT-5.5 Rosalind xhigh 55.9%
  21. 21 Gemini 3 Flash (Dec '25 Preview) high 55.6%
  22. 22 GLM-5.2 max 55.5%
  23. 23 Claude Opus 4.1 16k 55.1%
  24. 24 Gemini 3.7 Flash high 55%
  25. 25 GPT-5.2 xhigh 54.8%
  26. 26 Gemini 3.6 Flash high 54.3%
  27. 27 Claude Sonnet 4.5 16k 54.1%
  28. 28 DeepSeek V4.1 Flash 1017 max 53.9%
  29. 29 DeepSeek V4 Pro 0813 max 53%
  30. 30 Qwen 3.8 27B xhigh 53%
  31. 31 Claude Opus 4.5 16k 52.9%
  32. 32 Qwen3.5 397B-A17B 52.5%
  33. 33 Nemotron 3.5 Super VL 52.5%
  34. 34 Gemini 2.5 Pro May 2025, 16k 52.2%
  35. 35 Qwen3.7-Max 51.8%
  36. 36 Grok 4 51.7%
  37. 37 GPT-5 mini high 51.5%
  38. 38 Mistral Large 4 high 51.3%
  39. 39 Inkling xhigh 51.2%
  40. 40 o1 high 51.1%
  41. 41 GPT-5 high 50.8%
  42. 42 Grok 4.6 xhigh 50.4%
  43. 43 GLM-5.3 max 49.8%
  44. 44 GPT-5 Pro high 49.6%
  45. 45 o3 high 49.2%
  46. 46 GPT-5.1 high 49.2%
  47. 47 Beam 1.0 max 49.2%
  48. 48 o4-mini high 48.9%
  49. 49 Gemini 3.8 Flash high 48.9%
  50. 50 Gemini 3 Pro (Nov '25 Preview) high 48.7%
  51. 51 Gemma 4 31B IT 48.3%
  52. 52 Inkling-Small xhigh 48.2%
  53. 53 GPT-5.4 xhigh 48.1%
  54. 54 GLM-5 48%
  55. 55 Gemini 3.1 Pro (Feb 26' Preview) high 47.5%
  56. 56 Qwen3.6 27B 47.5%
  57. 57 Qwen 3.6 Plus 47.2%
  58. 58 Claude Sonnet 3.7 16k 46.8%
  59. 59 GPT-4.5 46.5%
  60. 60 Claude Sonnet 4 16k 46.4%
  61. 61 gpt-oss-120b 46.4%
  62. 62 Kimi K2 Thinking 46.1%
  63. 63 Grok 4.7 xhigh 46.1%
  64. 64 o3-mini high 45.7%
  65. 65 Mistral Medium 3.5 high 45.6%
  66. 66 Grok 4.1 Fast reasoning 45.1%
  67. 67 Gemini 3.5 Flash high 45%
  68. 68 Nemotron 3 Ultra 550B A55B 44.8%
  69. 69 Qwen3-235B-A22B-Thinking Jul 2025 44.4%
  70. 70 Sonar Deep Research 44.3%
  71. 71 Grok 4 Fast reasoning 44.3%
  72. 72 Nemotron 3 Super 120B A12B 43.9%
  73. 73 DeepSeek-R1 May 2025 43.1%
  74. 74 Claude 3.5 Sonnet (October 2024) 42.9%
  75. 75 GPT-4.1 42.5%
  76. 76 GPT-4.1 mini 42.5%
  77. 77 Claude Haiku 4.5 16k 42.5%
  78. 78 Kimi K2.5 42.4%
  79. 79 Qwen3-Next-80B-A3B thinking 42.2%
  80. 80 Qwen3-235B-A22B 41.4%
  81. 81 Grok 3 41%
  82. 82 GPT-5 nano high 41%
  83. 83 Kimi K2 (Jul '25) 40.7%
  84. 84 Kimi K2 (Sep '25) 40.4%
  85. 85 Llama 4 Maverick 40.2%
  86. 86 GPT-4o Nov 2024 39.9%
  87. 87 Gemini 2.5 Flash (Apr '25 Preview) 16k 39.8%
  88. 88 Gemini 2.5 Flash Jun 2025, 16k 39.8%
  89. 89 GPT-4o Aug 2024 39.6%
  90. 90 DeepSeek-V3.1 39.1%
  91. 91 Qwen3-Next-80B-A3B-Instruct 39.1%
  92. 92 DeepSeek-V3.2 39%
  93. 93 Nemotron 3 Nano 30B-A3B 38.5%
  94. 94 Grok-3 mini high 37.5%
  95. 95 Gemini 2.5 Pro (May '25 Preview) 16k 35.1%
  96. 96 DeepSeek-V4-Pro max 33.6%
  97. 97 GPT-4 Turbo Nov 2023 33.4%
  98. 98 Magistral Medium 1.0 30.8%
  99. 99 Claude Haiku 3.5 28%
  100. 100 GPT-4 Mar 2023 27.5%
  101. 101 Nemotron 3.5 Lightning 30B-A3B 26.1%
  102. 102 GPT-4.1 nano 24.5%
  103. 103 Magistral Small 24.4%
  104. 104 GPT-4o mini 22.4%
  105. 105 Muse Glimmer 30B xhigh 15.5%
  106. 106 GPT-3.5 Turbo Jun 2023 11.5%

Source: Official Molecular Biology Capabilities Test leaderboard

95 models. Scroll sideways for the rest.

  1. 68.7 Claude Opus 5.5
  2. 63.9 GPT-6 Astra
  3. 63.2 Gemini 4 Argon
  4. 61.1 GPT-5.6 Sol
  5. 61.1 GPT-6 Sol
  6. 60.7 Kimi K3.1 Pro
  7. 60.2 Claude Opus 5
  8. 59.8 Kimi K2.6
  9. 59.6 Muse Wonder Full
  10. 59.2 Claude Sonnet 5
  11. 59.1 Claude Opus 4
  12. 58.8 Claude Sonnet 5.5
  13. 58.4 Kimi K3
  14. 57.8 Qwen3.8-2.4T-A95B
  15. 57.1 GPT-5.5
  16. 56.9 GPT-5.6 Terra
  17. 56.7 Claude Opus 4.6
  18. 56.4 Claude Opus 4.8
  19. 56.0 Claude Opus 4.7
  20. 55.5 GLM-5.2
  21. 55.1 Claude Opus 4.1
  22. 55.0 Gemini 3.7 Flash
  23. 54.8 GPT-5.2
  24. 54.4 Muse Spark
  25. 54.3 Gemini 3.6 Flash
  26. 54.1 Claude Sonnet 4.5
  27. 53.9 DeepSeek V4.1 Flash 1017
  28. 53.2 Muse Spark 1.1
  29. 53.0 DeepSeek V4 Pro 0813
  30. 53.0 Qwen 3.8 27B
  31. 52.9 Claude Opus 4.5
  32. 52.5 Qwen3.5 397B-A17B
  33. 52.5 Nemotron 3.5 Super VL
  34. 52.2 Gemini 2.5 Pro
  35. 52.2 Muse Spark 1.3
  36. 51.8 Qwen3.7-Max
  37. 51.7 Grok 4
  38. 51.5 GPT-5 mini
  39. 51.3 Mistral Large 4
  40. 51.2 Inkling
  41. 51.1 o1
  42. 50.8 GPT-5
  43. 50.4 Grok 4.6
  44. 49.8 GLM-5.3
  45. 49.6 GPT-5 Pro
  46. 49.2 o3
  47. 49.2 GPT-5.1
  48. 49.2 Beam 1.0
  49. 48.9 o4-mini
  50. 48.9 Gemini 3.8 Flash
  51. 48.3 Gemma 4 31B IT
  52. 48.2 Inkling-Small
  53. 48.1 GPT-5.4
  54. 48.0 GLM-5
  55. 47.5 Qwen3.6 27B
  56. 47.2 Qwen 3.6 Plus
  57. 46.5 GPT-4.5
  58. 46.4 Claude Sonnet 4
  59. 46.4 gpt-oss-120b
  60. 46.1 Kimi K2 Thinking
  61. 46.1 Grok 4.7
  62. 45.7 o3-mini
  63. 45.6 Mistral Medium 3.5
  64. 45.1 Grok 4.1 Fast
  65. 45.0 Gemini 3.5 Flash
  66. 44.4 Qwen3-235B-A22B-Thinking
  67. 44.3 Grok 4 Fast
  68. 43.1 DeepSeek-R1
  69. 42.9 Claude 3.5 Sonnet (October 2024)
  70. 42.5 GPT-4.1
  71. 42.5 GPT-4.1 mini
  72. 42.5 Claude Haiku 4.5
  73. 42.4 Kimi K2.5
  74. 42.2 Qwen3-Next-80B-A3B
  75. 41.4 Qwen3-235B-A22B
  76. 41.0 Grok 3
  77. 41.0 GPT-5 nano
  78. 40.2 Llama 4 Maverick
  79. 39.9 GPT-4o
  80. 39.8 Gemini 2.5 Flash
  81. 39.6 GPT-4o
  82. 39.1 DeepSeek-V3.1
  83. 39.1 Qwen3-Next-80B-A3B-Instruct
  84. 39.0 DeepSeek-V3.2
  85. 38.5 Nemotron 3 Nano 30B-A3B
  86. 37.5 Grok-3 mini
  87. 33.6 DeepSeek-V4-Pro
  88. 33.4 GPT-4 Turbo
  89. 32.7 Code World Model (CWM)
  90. 30.8 Magistral Medium 1.0
  91. 27.5 GPT-4
  92. 26.1 Nemotron 3.5 Lightning 30B-A3B
  93. 24.5 GPT-4.1 nano
  94. 22.4 GPT-4o mini
  95. 11.5 GPT-3.5 Turbo
  1. 1 Claude Opus 5.5 max 68.7%
  2. 2 GPT-6 Astra max 63.9%
  3. 3 Gemini 4 Argon high 63.2%
  4. 4 GPT-5.6 Sol max 61.1%
  5. 5 GPT-6 Sol max 61.1%
  6. 6 Kimi K3.1 Pro max 60.7%
  7. 7 Claude Opus 5 max 60.2%
  8. 8 Kimi K2.6 59.8%
  9. 9 Muse Wonder Full 59.6%
  10. 10 Claude Sonnet 5 max 59.2%
  11. 11 Claude Opus 4 16k 59.1%
  12. 12 Claude Sonnet 5.5 max 58.8%
  13. 13 Kimi K3 max 58.4%
  14. 14 Qwen3.8-2.4T-A95B xhigh 57.8%
  15. 15 GPT-5.5 xhigh 57.1%
  16. 16 GPT-5.6 Terra max 56.9%
  17. 17 Claude Opus 4.6 max 56.7%
  18. 18 Claude Opus 4.8 max 56.4%
  19. 19 Claude Opus 4.7 max 56%
  20. 20 GLM-5.2 max 55.5%
  21. 21 Claude Opus 4.1 16k 55.1%
  22. 22 Gemini 3.7 Flash high 55%
  23. 23 GPT-5.2 xhigh 54.8%
  24. 24 Muse Spark 54.4%
  25. 25 Gemini 3.6 Flash high 54.3%
  26. 26 Claude Sonnet 4.5 16k 54.1%
  27. 27 DeepSeek V4.1 Flash 1017 max 53.9%
  28. 28 Muse Spark 1.1 53.2%
  29. 29 DeepSeek V4 Pro 0813 max 53%
  30. 30 Qwen 3.8 27B xhigh 53%
  31. 31 Claude Opus 4.5 16k 52.9%
  32. 32 Qwen3.5 397B-A17B 52.5%
  33. 33 Nemotron 3.5 Super VL 52.5%
  34. 34 Gemini 2.5 Pro May 2025, 16k 52.2%
  35. 35 Muse Spark 1.3 52.2%
  36. 36 Qwen3.7-Max 51.8%
  37. 37 Grok 4 51.7%
  38. 38 GPT-5 mini high 51.5%
  39. 39 Mistral Large 4 high 51.3%
  40. 40 Inkling xhigh 51.2%
  41. 41 o1 high 51.1%
  42. 42 GPT-5 high 50.8%
  43. 43 Grok 4.6 xhigh 50.4%
  44. 44 GLM-5.3 max 49.8%
  45. 45 GPT-5 Pro high 49.6%
  46. 46 o3 high 49.2%
  47. 47 GPT-5.1 high 49.2%
  48. 48 Beam 1.0 max 49.2%
  49. 49 o4-mini high 48.9%
  50. 50 Gemini 3.8 Flash high 48.9%
  51. 51 Gemma 4 31B IT 48.3%
  52. 52 Inkling-Small xhigh 48.2%
  53. 53 GPT-5.4 xhigh 48.1%
  54. 54 GLM-5 48%
  55. 55 Qwen3.6 27B 47.5%
  56. 56 Qwen 3.6 Plus 47.2%
  57. 57 GPT-4.5 46.5%
  58. 58 Claude Sonnet 4 16k 46.4%
  59. 59 gpt-oss-120b 46.4%
  60. 60 Kimi K2 Thinking 46.1%
  61. 61 Grok 4.7 xhigh 46.1%
  62. 62 o3-mini high 45.7%
  63. 63 Mistral Medium 3.5 high 45.6%
  64. 64 Grok 4.1 Fast reasoning 45.1%
  65. 65 Gemini 3.5 Flash high 45%
  66. 66 Qwen3-235B-A22B-Thinking Jul 2025 44.4%
  67. 67 Grok 4 Fast reasoning 44.3%
  68. 68 DeepSeek-R1 May 2025 43.1%
  69. 69 Claude 3.5 Sonnet (October 2024) 42.9%
  70. 70 GPT-4.1 42.5%
  71. 71 GPT-4.1 mini 42.5%
  72. 72 Claude Haiku 4.5 16k 42.5%
  73. 73 Kimi K2.5 42.4%
  74. 74 Qwen3-Next-80B-A3B thinking 42.2%
  75. 75 Qwen3-235B-A22B 41.4%
  76. 76 Grok 3 41%
  77. 77 GPT-5 nano high 41%
  78. 78 Llama 4 Maverick 40.2%
  79. 79 GPT-4o Nov 2024 39.9%
  80. 80 Gemini 2.5 Flash Jun 2025, 16k 39.8%
  81. 81 GPT-4o Aug 2024 39.6%
  82. 82 DeepSeek-V3.1 39.1%
  83. 83 Qwen3-Next-80B-A3B-Instruct 39.1%
  84. 84 DeepSeek-V3.2 39%
  85. 85 Nemotron 3 Nano 30B-A3B 38.5%
  86. 86 Grok-3 mini high 37.5%
  87. 87 DeepSeek-V4-Pro max 33.6%
  88. 88 GPT-4 Turbo Nov 2023 33.4%
  89. 89 Code World Model (CWM) 32.7%
  90. 90 Magistral Medium 1.0 30.8%
  91. 91 GPT-4 Mar 2023 27.5%
  92. 92 Nemotron 3.5 Lightning 30B-A3B 26.1%
  93. 93 GPT-4.1 nano 24.5%
  94. 94 GPT-4o mini 22.4%
  95. 95 GPT-3.5 Turbo Jun 2023 11.5%

Sources: Official Molecular Biology Capabilities Test leaderboard and labs’ model cards

#EntryScoreSettingDateSource
1 Claude Opus 5.5 68.7% max Sep 2026 Official leaderboard
2 GPT-6 Astra 63.9% max Sep 2026 Official leaderboard
3 Gemini 4 Argon 63.2% high Sep 2026 Official leaderboard
4 GPT-5.6 Sol 61.1% max Jul 2026 Official leaderboard
5 GPT-6 Sol 61.1% max Sep 2026 Official leaderboard
6 Kimi K3.1 Pro 60.7% max Oct 2026 Official leaderboard
7 Claude Opus 5 60.2% max Jul 2026 Official leaderboard
8 Kimi K2.6 59.8% Apr 2026 Official leaderboard
9 Claude Sonnet 5 59.2% max Jun 2026 Official leaderboard
10 Claude Opus 4 59.1% 16k May 2025 Official leaderboard
11 Claude Sonnet 5.5 58.8% max Sep 2026 Official leaderboard
12 Kimi K3 58.4% max Jul 2026 Official leaderboard
13 Qwen3.8-2.4T-A95B 57.8% xhigh Aug 2026 Official leaderboard
14 GPT-5.5 57.1% xhigh Apr 2026 Official leaderboard
15 GPT-5.5 pro Codex 57% xhigh Apr 2026 Official leaderboard
16 GPT-5.6 Terra 56.9% max Jul 2026 Official leaderboard
17 Claude Opus 4.6 56.7% max Feb 2026 Official leaderboard
18 Claude Opus 4.8 56.4% max May 2026 Official leaderboard
19 Claude Opus 4.7 56% max Apr 2026 Official leaderboard
20 GPT-5.5 Rosalind 55.9% xhigh Jun 2026 Official leaderboard
21 Gemini 3 Flash (Dec '25 Preview) 55.6% high Dec 2025 Official leaderboard
22 GLM-5.2 55.5% max Jun 2026 Official leaderboard
23 Claude Opus 4.1 55.1% 16k Aug 2025 Official leaderboard
24 Gemini 3.7 Flash 55% high Aug 2026 Official leaderboard
25 GPT-5.2 54.8% xhigh Dec 2025 Official leaderboard
26 Gemini 3.6 Flash 54.3% high Jul 2026 Official leaderboard
27 Claude Sonnet 4.5 54.1% 16k Sep 2025 Official leaderboard
28 DeepSeek V4.1 Flash 1017 53.9% max Oct 2026 Official leaderboard
29 DeepSeek V4 Pro 0813 53% max Aug 2026 Official leaderboard
30 Qwen 3.8 27B 53% xhigh Aug 2026 Official leaderboard
31 Claude Opus 4.5 52.9% 16k Nov 2025 Official leaderboard
32 Qwen3.5 397B-A17B 52.5% Feb 2026 Official leaderboard
33 Nemotron 3.5 Super VL 52.5% Oct 2026 Official leaderboard
34 Gemini 2.5 Pro May 2025 52.2% 16k May 2025 Official leaderboard
35 Qwen3.7-Max 51.8% May 2026 Official leaderboard
36 Grok 4 51.7% Jul 2025 Official leaderboard
37 GPT-5 mini 51.5% high Aug 2025 Official leaderboard
38 Mistral Large 4 51.3% high Oct 2026 Official leaderboard
39 Inkling 51.2% xhigh Jul 2026 Official leaderboard
40 o1 51.1% high Dec 2024 Official leaderboard
41 GPT-5 50.8% high Aug 2025 Official leaderboard
42 Grok 4.6 50.4% xhigh Aug 2026 Official leaderboard
43 GLM-5.3 49.8% max Aug 2026 Official leaderboard
44 GPT-5 Pro 49.6% high Oct 2025 Official leaderboard
45 o3 49.2% high Apr 2025 Official leaderboard
46 GPT-5.1 49.2% high Nov 2025 Official leaderboard
47 Beam 1.0 49.2% max Oct 2026 Official leaderboard
48 o4-mini 48.9% high Apr 2025 Official leaderboard
49 Gemini 3.8 Flash 48.9% high Sep 2026 Official leaderboard
50 Gemini 3 Pro (Nov '25 Preview) 48.7% high Nov 2025 Official leaderboard
51 Gemma 4 31B IT 48.3% Apr 2026 Official leaderboard
52 Inkling-Small 48.2% xhigh Jul 2026 Official leaderboard
53 GPT-5.4 48.1% xhigh Mar 2026 Official leaderboard
54 GLM-5 48% Feb 2026 Official leaderboard
55 Gemini 3.1 Pro (Feb 26' Preview) 47.5% high Feb 2026 Official leaderboard
56 Qwen3.6 27B 47.5% Apr 2026 Official leaderboard
57 Qwen 3.6 Plus 47.2% Mar 2026 Official leaderboard
58 Claude Sonnet 3.7 46.8% 16k Feb 2025 Official leaderboard
59 GPT-4.5 46.5% Feb 2025 Official leaderboard
60 Claude Sonnet 4 46.4% 16k May 2025 Official leaderboard
61 gpt-oss-120b 46.4% Aug 2025 Official leaderboard
62 Kimi K2 Thinking 46.1% Nov 2025 Official leaderboard
63 Grok 4.7 46.1% xhigh Sep 2026 Official leaderboard
64 o3-mini 45.7% high Jan 2025 Official leaderboard
65 Mistral Medium 3.5 45.6% high Apr 2026 Official leaderboard
66 Grok 4.1 Fast 45.1% reasoning Nov 2025 Official leaderboard
67 Gemini 3.5 Flash 45% high May 2026 Official leaderboard
68 Nemotron 3 Ultra 550B A55B 44.8% Jun 2026 Official leaderboard
69 Qwen3-235B-A22B-Thinking Jul 2025 44.4% Jul 2025 Official leaderboard
70 Sonar Deep Research 44.3% Feb 2025 Official leaderboard
71 Grok 4 Fast 44.3% reasoning Sep 2025 Official leaderboard
72 Nemotron 3 Super 120B A12B 43.9% Mar 2026 Official leaderboard
73 DeepSeek-R1 May 2025 43.1% May 2025 Official leaderboard
74 Claude 3.5 Sonnet (October 2024) 42.9% Oct 2024 Official leaderboard
75 GPT-4.1 42.5% Apr 2025 Official leaderboard
76 GPT-4.1 mini 42.5% Apr 2025 Official leaderboard
77 Claude Haiku 4.5 42.5% 16k Oct 2025 Official leaderboard
78 Kimi K2.5 42.4% Jan 2026 Official leaderboard
79 Qwen3-Next-80B-A3B 42.2% thinking Sep 2025 Official leaderboard
80 Qwen3-235B-A22B 41.4% Apr 2025 Official leaderboard
81 Grok 3 41% Apr 2025 Official leaderboard
82 GPT-5 nano 41% high Aug 2025 Official leaderboard
83 Kimi K2 (Jul '25) 40.7% Jul 2025 Official leaderboard
84 Kimi K2 (Sep '25) 40.4% Sep 2025 Official leaderboard
85 Llama 4 Maverick 40.2% Apr 2025 Official leaderboard
86 GPT-4o Nov 2024 39.9% Nov 2024 Official leaderboard
87 Gemini 2.5 Flash (Apr '25 Preview) 39.8% 16k Apr 2025 Official leaderboard
88 Gemini 2.5 Flash Jun 2025 39.8% 16k Jun 2025 Official leaderboard
89 GPT-4o Aug 2024 39.6% Aug 2024 Official leaderboard
90 DeepSeek-V3.1 39.1% Aug 2025 Official leaderboard
91 Qwen3-Next-80B-A3B-Instruct 39.1% Sep 2025 Official leaderboard
92 DeepSeek-V3.2 39% Dec 2025 Official leaderboard
93 Nemotron 3 Nano 30B-A3B 38.5% Dec 2025 Official leaderboard
94 Grok-3 mini 37.5% high Apr 2025 Official leaderboard
95 Gemini 2.5 Pro (May '25 Preview) 35.1% 16k May 2025 Official leaderboard
96 DeepSeek-V4-Pro 33.6% max Apr 2026 Official leaderboard
97 GPT-4 Turbo Nov 2023 33.4% Nov 2023 Official leaderboard
98 Magistral Medium 1.0 30.8% Jun 2025 Official leaderboard
99 Claude Haiku 3.5 28% Oct 2024 Official leaderboard
100 GPT-4 Mar 2023 27.5% Mar 2023 Official leaderboard
101 Nemotron 3.5 Lightning 30B-A3B 26.1% Aug 2026 Official leaderboard
102 GPT-4.1 nano 24.5% Apr 2025 Official leaderboard
103 Magistral Small 24.4% Jun 2025 Official leaderboard
104 GPT-4o mini 22.4% Jul 2024 Official leaderboard
105 Muse Glimmer 30B 15.5% xhigh Aug 2026 Official leaderboard
106 GPT-3.5 Turbo Jun 2023 11.5% Jun 2023 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 5.5 68.7% max Sep 2026 Official leaderboard
2 GPT-6 Astra 63.9% max Sep 2026 Official leaderboard
3 Gemini 4 Argon 63.2% high Sep 2026 Official leaderboard
4 GPT-5.6 Sol 61.1% max Jul 2026 Official leaderboard
5 GPT-6 Sol 61.1% max Sep 2026 Official leaderboard
6 Kimi K3.1 Pro 60.7% max Oct 2026 Official leaderboard
7 Claude Opus 5 60.2% max Jul 2026 Official leaderboard
8 Kimi K2.6 59.8% Apr 2026 Official leaderboard
9 Muse Wonder Full 59.6% Oct 2026 Lab model card
10 Claude Sonnet 5 59.2% max Jun 2026 Official leaderboard
11 Claude Opus 4 59.1% 16k May 2025 Official leaderboard
12 Claude Sonnet 5.5 58.8% max Sep 2026 Official leaderboard
13 Kimi K3 58.4% max Jul 2026 Official leaderboard
14 Qwen3.8-2.4T-A95B 57.8% xhigh Aug 2026 Official leaderboard
15 GPT-5.5 57.1% xhigh Apr 2026 Official leaderboard
16 GPT-5.6 Terra 56.9% max Jul 2026 Official leaderboard
17 Claude Opus 4.6 56.7% max Feb 2026 Official leaderboard
18 Claude Opus 4.8 56.4% max May 2026 Official leaderboard
19 Claude Opus 4.7 56% max Apr 2026 Official leaderboard
20 GLM-5.2 55.5% max Jun 2026 Official leaderboard
21 Claude Opus 4.1 55.1% 16k Aug 2025 Official leaderboard
22 Gemini 3.7 Flash 55% high Aug 2026 Official leaderboard
23 GPT-5.2 54.8% xhigh Dec 2025 Official leaderboard
24 Muse Spark 54.4% Jun 2026 Lab model card
25 Gemini 3.6 Flash 54.3% high Jul 2026 Official leaderboard
26 Claude Sonnet 4.5 54.1% 16k Sep 2025 Official leaderboard
27 DeepSeek V4.1 Flash 1017 53.9% max Oct 2026 Official leaderboard
28 Muse Spark 1.1 53.2% Jul 2026 Lab model card
29 DeepSeek V4 Pro 0813 53% max Aug 2026 Official leaderboard
30 Qwen 3.8 27B 53% xhigh Aug 2026 Official leaderboard
31 Claude Opus 4.5 52.9% 16k Nov 2025 Official leaderboard
32 Qwen3.5 397B-A17B 52.5% Feb 2026 Official leaderboard
33 Nemotron 3.5 Super VL 52.5% Oct 2026 Official leaderboard
34 Gemini 2.5 Pro May 2025 52.2% 16k May 2025 Official leaderboard
35 Muse Spark 1.3 52.2% Oct 2026 Lab model card
36 Qwen3.7-Max 51.8% May 2026 Official leaderboard
37 Grok 4 51.7% Jul 2025 Official leaderboard
38 GPT-5 mini 51.5% high Aug 2025 Official leaderboard
39 Mistral Large 4 51.3% high Oct 2026 Official leaderboard
40 Inkling 51.2% xhigh Jul 2026 Official leaderboard
41 o1 51.1% high Dec 2024 Official leaderboard
42 GPT-5 50.8% high Aug 2025 Official leaderboard
43 Grok 4.6 50.4% xhigh Aug 2026 Official leaderboard
44 GLM-5.3 49.8% max Aug 2026 Official leaderboard
45 GPT-5 Pro 49.6% high Oct 2025 Official leaderboard
46 o3 49.2% high Apr 2025 Official leaderboard
47 GPT-5.1 49.2% high Nov 2025 Official leaderboard
48 Beam 1.0 49.2% max Oct 2026 Official leaderboard
49 o4-mini 48.9% high Apr 2025 Official leaderboard
50 Gemini 3.8 Flash 48.9% high Sep 2026 Official leaderboard
51 Gemma 4 31B IT 48.3% Apr 2026 Official leaderboard
52 Inkling-Small 48.2% xhigh Jul 2026 Official leaderboard
53 GPT-5.4 48.1% xhigh Mar 2026 Official leaderboard
54 GLM-5 48% Feb 2026 Official leaderboard
55 Qwen3.6 27B 47.5% Apr 2026 Official leaderboard
56 Qwen 3.6 Plus 47.2% Mar 2026 Official leaderboard
57 GPT-4.5 46.5% Feb 2025 Official leaderboard
58 Claude Sonnet 4 46.4% 16k May 2025 Official leaderboard
59 gpt-oss-120b 46.4% Aug 2025 Official leaderboard
60 Kimi K2 Thinking 46.1% Nov 2025 Official leaderboard
61 Grok 4.7 46.1% xhigh Sep 2026 Official leaderboard
62 o3-mini 45.7% high Jan 2025 Official leaderboard
63 Mistral Medium 3.5 45.6% high Apr 2026 Official leaderboard
64 Grok 4.1 Fast 45.1% reasoning Nov 2025 Official leaderboard
65 Gemini 3.5 Flash 45% high May 2026 Official leaderboard
66 Qwen3-235B-A22B-Thinking Jul 2025 44.4% Jul 2025 Official leaderboard
67 Grok 4 Fast 44.3% reasoning Sep 2025 Official leaderboard
68 DeepSeek-R1 May 2025 43.1% May 2025 Official leaderboard
69 Claude 3.5 Sonnet (October 2024) 42.9% Oct 2024 Official leaderboard
70 GPT-4.1 42.5% Apr 2025 Official leaderboard
71 GPT-4.1 mini 42.5% Apr 2025 Official leaderboard
72 Claude Haiku 4.5 42.5% 16k Oct 2025 Official leaderboard
73 Kimi K2.5 42.4% Jan 2026 Official leaderboard
74 Qwen3-Next-80B-A3B 42.2% thinking Sep 2025 Official leaderboard
75 Qwen3-235B-A22B 41.4% Apr 2025 Official leaderboard
76 Grok 3 41% Apr 2025 Official leaderboard
77 GPT-5 nano 41% high Aug 2025 Official leaderboard
78 Llama 4 Maverick 40.2% Apr 2025 Official leaderboard
79 GPT-4o Nov 2024 39.9% Nov 2024 Official leaderboard
80 Gemini 2.5 Flash Jun 2025 39.8% 16k Jun 2025 Official leaderboard
81 GPT-4o Aug 2024 39.6% Aug 2024 Official leaderboard
82 DeepSeek-V3.1 39.1% Aug 2025 Official leaderboard
83 Qwen3-Next-80B-A3B-Instruct 39.1% Sep 2025 Official leaderboard
84 DeepSeek-V3.2 39% Dec 2025 Official leaderboard
85 Nemotron 3 Nano 30B-A3B 38.5% Dec 2025 Official leaderboard
86 Grok-3 mini 37.5% high Apr 2025 Official leaderboard
87 DeepSeek-V4-Pro 33.6% max Apr 2026 Official leaderboard
88 GPT-4 Turbo Nov 2023 33.4% Nov 2023 Official leaderboard
89 Code World Model (CWM) 32.7% Oct 2025 Lab model card
90 Magistral Medium 1.0 30.8% Jun 2025 Official leaderboard
91 GPT-4 Mar 2023 27.5% Mar 2023 Official leaderboard
92 Nemotron 3.5 Lightning 30B-A3B 26.1% Aug 2026 Official leaderboard
93 GPT-4.1 nano 24.5% Apr 2025 Official leaderboard
94 GPT-4o mini 22.4% Jul 2024 Official leaderboard
95 GPT-3.5 Turbo Jun 2023 11.5% Jun 2023 Official leaderboard

Labs report scores for 8 models in their cards. Switch the scores above to rank them together.

11 models have held number one on the official leaderboard since 2023 11 models have held number one with claims included since 2023

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 what labs claim in their cards.

OpenAI Anthropic Google DeepMind Moonshot AI Meta
Own leaderboard 11 leaders
Claude Opus 5.5 68.7%
Lab claims 3 leaders
GPT-5.6 Sol 60%

Sources: Official Molecular Biology Capabilities Test leaderboard and labs’ model cards

Scores over time

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 Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 2027 Near the ceiling (85%) Claude Opus 5.5 GPT-6 Astra Claude Opus 4 Gemini 2.5 Pro Claude 3.5 Sonnet (October 2024) GPT-4 Turbo GPT-4 0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 2027 Near the ceiling (85%) Claude Opus 5.5 GPT-5.6 Sol

Sources: Official Molecular Biology Capabilities Test leaderboard and labs’ model cards

Labs report it only in system cards

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.

9 system cards report it, from 3 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
Oct 23, 2026 Muse Spark 1.3 system card 52.2 MBCT accuracy, Muse Spark 1.3 (API) max, multiple-response MC (Table 1, Figure 1, 2.1.1.1) Bio and chem risk
Oct 12, 2026 Muse Wonder Full model card 59.6 SecureBio MBCT, Muse Wonder Full (CBRN Capability Evaluations table, section 4). SecureBio average of MBCT, HPCT and VCT 60.7 Bio and chem risk
Oct 5, 2026 Beam 1.0 technical report 48.5 MBCT exact-match accuracy Bio and chem risk
Jul 9, 2026 GPT-5.6 Sol system card 60 SecureBio Molecular Biology Capabilities Test, strongest reported configuration (9.1.1.8) Bio and chem risk
Jul 9, 2026 Muse Spark 1.1 system card 53.2 MBCT, Muse Spark 1.1 Helpful (Table 1, Figure 1) Bio and chem risk
Jun 25, 2026 GPT-5.6 Sol system card 60 SecureBio Molecular Biology Capabilities Test, strongest reported configuration (9.1.1.8) Bio and chem risk
Jun 12, 2026 Muse Spark system card 54.4 MBCT, Muse Spark Thinking (Table 1, Figure 1) Bio and chem risk
Oct 8, 2025 Code World Model (CWM) technical report 32.7 Molecular Biology Capabilities Test (MBCT) (Table 9) Bio and chem risk
Aug 7, 2025 GPT-5 system card 48 SecureBio Molecular Biology Capabilities Test, mean of n=10 (Table 15) Bio and chem risk

The lab's number comes out ahead in 0 of 3 pairs that differ

Each score a lab reported for its own model, beside the same model's best score on the leaderboard. Gaps mostly come from different harnesses, effort levels and task subsets. Every benchmark.

Model Lab says Board says Gap
GPT-5 48% SecureBio Molecular Biology Capabilities Test, mean of n=10 (Table 15) 50.8% Leaderboard, high −2.8 pts
GPT-5.6 Sol 60% SecureBio Molecular Biology Capabilities Test, strongest reported configuration (9.1.1.8) 61.1% Leaderboard, max −1.1 pts
Beam 1.0 48.5% MBCT exact-match accuracy 49.2% Leaderboard, max −0.7 pts

More in Biology and chemistry

All 108 in Biology and chemistry
  • GeneBench Pro In use Agents analyze messy simulated genomics and biology datasets to estimate a decision-relevant quantity across 129 problems. Scored automatically against known answers. GPT-5.6 Sol 31.5%
  • SUPERChem In use Models solve 500 expert-written chemistry reasoning problems, offered in multimodal and text-only forms. Expert solution paths also allow a reasoning path score. Tencent Hy4 preview 66.4%
  • LifeSciBench In use Models complete 750 expert-written life science research tasks, often with attached figures, tables, or sequence files. Graded against expert rubrics as score and pass rate. GPT-Rosalind-6 61.3%
  • MedChemBench (OpenAI) In use Models answer chemistry questions from an OpenAI internal set, grouped with OpenAI's life science research evaluations. OpenAI has not published the task format. GPT-Rosalind-6 49.6%
  • ProtocolQA Open-Ended In use Models correct errors in published lab protocols using open-ended questions adapted from LAB-Bench ProtocolQA. Scored as pass@1. GPT-6 Cyber 47.2%
  • MolecularIQ In use Models answer questions about molecular graphs, such as counting, indexing and constrained generation, and a symbolic checker verifies each answer. gpt-oss-120b 47.5%