Benchmark Almanac

Tax Agent Bench

Gemini 4 Hydrogen 52.3% leads since November 2026
Gemini 4 Hydrogen 52.3% leads since November 2026
In use

The official leaderboard hasn’t scored a model released after GPT-6.2 Sol (November 2026). Labs report their own numbers for one newer model: Inkling 2 Mini at 14.1%.

Agents research 391 expert-written US corporate tax questions with tools and submit one cited written answer. A judge grades it against a rubric and checks that citations are real.

Splits are 5 public, 193 licensable validation and 193 private test questions. Main metric is all-pass accuracy, with weighted pass as partial credit. 80 models on the board on October 29th, 2026.

Made by
Vals AI
Released
September 2026
Kind
Finance
Level
Expert
Scored
Graded by a model
Answers
Agent task
Tasks
Public, plus a private test set
Measures
Capability
Language
English
In and out
Text → Text, Actions
Interaction
Agentic
Environment
Browser
Items from
Written by experts
Refresh
One fixed set
Run by
A host
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last in a launch post
November 2026
Last model added
November 2026
Scores refreshed
November 18th, 2026
88
Entries on its own leaderboard, 87 models we track
2
Launch posts list it, from 2 labs
1
System card reports it, from 1 lab

Gemini 4 Hydrogen leads with 52.3% Gemini 4 Hydrogen leads with 52.3%

Best score per entry from the benchmark's own leaderboard, read on November 18th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. 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.

Gemini 4 Hydrogen leads with 52.3%, 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. Striped bars are claims, and dotted bars are reruns by Epoch AI or another board.

88 entries. Scroll sideways for the rest.

  1. 52.3 Gemini 4 Hydrogen
  2. 51.8 Gemini 4 Carbon
  3. 50.8 Claude Fable 5.1
  4. 49.7 Qwen3.8 Max (0902)
  5. 49.2 Gemini 4 Argon
  6. 47.1 Claude Opus 5
  7. 44.5 Claude Opus 5.5
  8. 44.4 DeepSeek V4.1 Flash 1017
  9. 43.3 GLM-5.3
  10. 43.0 Muse Wonder Full
  11. 42.4 Muse Spark 1.3
  12. 42.0 Claude Sonnet 5.5
  13. 40.4 Qwen4-Max
  14. 38.9 Inkling 2 Small
  15. 38.3 Claude Fable 5
  16. 38.2 Muse Ember
  17. 36.3 Muse Wonder
  18. 35.8 Grok 4.7
  19. 34.4 Tencent Hy4
  20. 33.7 Grok 4.6
  21. 33.4 Kimi K3
  22. 32.8 Muse Spark 1.2
  23. 32.6 Tencent Hy4 preview
  24. 32.4 MiMo-V2.6 Pro
  25. 32.1 MiniMax-M3.1-Flash
  26. 32.1 Gemini 3.8 Flash
  27. 32.1 Qwen 3.8 Max
  28. 32.1 Claude Fable 5.2
  29. 31.9 Claude Haiku 5.5
  30. 31.6 GLM-5.3-Flash
  31. 31.1 GPT-5.6 Sol
  32. 31.1 Ember-1
  33. 30.9 Qwen 3.8 27B
  34. 30.7 Mistral Large 4
  35. 30.4 Mistral Large 4 Preview
  36. 30.0 GPT-5.6 Terra
  37. 30.0 Claude Opus 4.8
  38. 29.4 Claude Sonnet 5
  39. 28.5 MiMo-V2.6 Flash
  40. 28.1 DeepSeek V4 Flash 0731
  41. 27.3 Step 5 Preview
  42. 26.8 DeepSeek V4 Pro 0813
  43. 26.6 DeepSeek V4.1 Flash
  44. 26.3 Muse Spark Open
  45. 25.8 Claude Sonnet 4.6
  46. 25.2 DeepSeek-V4-Pro
  47. 24.9 Kimi K3.1 Pro
  48. 24.9 Grok 4.5
  49. 24.8 GPT-6.2 Sol
  50. 24.8 GPT-5.6 Luna
  51. 23.7 Claude Opus 4.7
  52. 23.3 Hy4 Flash
  53. 22.8 Gemini 3.7 Flash
  54. 21.8 GPT-6.2 Astra
  55. 21.8 GPT-5.5
  56. 21.8 GPT-6.1 Sol
  57. 21.6 Gemini 3.5 Flash
  58. 20.7 GPT-6 Astra
  59. 20.6 Ling-3.0-flash-Fin
  60. 20.6 MiniMax-M3
  61. 19.9 GPT-6 Luna
  62. 18.2 Mercury 3
  63. 18.0 Gemini 3.6 Flash
  64. 17.1 GPT-6.1 Terra
  65. 15.8 Qwen3.7-Max
  66. 15.4 Inkling-Small
  67. 15.3 Inkling
  68. 15.0 GPT-6 Sol
  69. 12.4 Kimi K2.6
  70. 11.8 Gemini 3 Flash
  71. 11.6 Nemotron 3.5 Super VL
  72. 11.5 MiMo-V2.5-Pro
  73. 10.5 Qwen3.7-Plus
  74. 9.29 Gemini 3.1 Pro
  75. 8.81 GPT-5.4 Mini
  76. 8.81 Beam 1.0
  77. 7.77 MiMo-V2.5
  78. 7.68 Qwen 3.6 Plus
  79. 7.25 Gemini 3.5 Flash-Lite
  80. 6.95 Grok 4.20
  81. 6.87 MiniMax-M2.7
  82. 6.18 Grok 4.3
  83. 5.86 Claude Haiku 4.5
  84. 4.97 GPT-5.4 Nano
  85. 4.48 Mistral Medium 3.5
  86. 3.11 Gemini 3.1 Flash-Lite
  87. 2.59 Nemotron 3.5 Lightning 30B-A3B
  88. 2.49 Mercury 2.5
  1. 1 Gemini 4 Hydrogen 52.3%
  2. 2 Gemini 4 Carbon 51.8%
  3. 3 Claude Fable 5.1 50.8%
  4. 4 Qwen3.8 Max (0902) 49.7%
  5. 5 Gemini 4 Argon 49.2%
  6. 6 Claude Opus 5 47.1%
  7. 7 Claude Opus 5.5 44.5%
  8. 8 DeepSeek V4.1 Flash 1017 44.4%
  9. 9 GLM-5.3 43.3%
  10. 10 Muse Wonder Full 43%
  11. 11 Muse Spark 1.3 max 42.4%
  12. 12 Claude Sonnet 5.5 42%
  13. 13 Qwen4-Max 40.4%
  14. 14 Inkling 2 Small 38.9%
  15. 15 Claude Fable 5 38.3%
  16. 16 Muse Ember 38.2%
  17. 17 Muse Wonder 36.3%
  18. 18 Grok 4.7 35.8%
  19. 19 Tencent Hy4 34.4%
  20. 20 Grok 4.6 33.7%
  21. 21 Kimi K3 33.4%
  22. 22 Muse Spark 1.2 32.8%
  23. 23 Tencent Hy4 preview 32.6%
  24. 24 MiMo-V2.6 Pro 32.4%
  25. 25 MiniMax-M3.1-Flash 32.1%
  26. 26 Gemini 3.8 Flash 32.1%
  27. 27 Qwen 3.8 Max 32.1%
  28. 28 Claude Fable 5.2 32.1%
  29. 29 Claude Haiku 5.5 31.9%
  30. 30 GLM-5.3-Flash 31.6%
  31. 31 GPT-5.6 Sol 31.1%
  32. 32 Ember-1 31.1%
  33. 33 Qwen 3.8 27B 30.9%
  34. 34 Mistral Large 4 30.7%
  35. 35 Mistral Large 4 Preview 30.4%
  36. 36 GPT-5.6 Terra 30%
  37. 37 Claude Opus 4.8 30%
  38. 38 Claude Sonnet 5 29.4%
  39. 39 MiMo-V2.6 Flash 28.5%
  40. 40 DeepSeek V4 Flash 0731 28.1%
  41. 41 Step 5 Preview 27.3%
  42. 42 DeepSeek V4 Pro 0813 26.8%
  43. 43 DeepSeek V4.1 Flash 26.6%
  44. 44 Muse Spark Open 26.3%
  45. 45 Claude Sonnet 4.6 25.8%
  46. 46 DeepSeek-V4-Pro 25.2%
  47. 47 Kimi K3.1 Pro 24.9%
  48. 48 Grok 4.5 24.9%
  49. 49 GPT-6.2 Sol 24.8%
  50. 50 GPT-5.6 Luna 24.8%
  51. 51 Claude Opus 4.7 23.7%
  52. 52 Hy4 Flash 23.3%
  53. 53 Gemini 3.7 Flash 22.8%
  54. 54 GPT-6.2 Astra 21.8%
  55. 55 GPT-5.5 21.8%
  56. 56 GPT-6.1 Sol 21.8%
  57. 57 Gemini 3.5 Flash 21.6%
  58. 58 GPT-6 Astra 20.7%
  59. 59 Ling-3.0-flash-Fin 20.6%
  60. 60 MiniMax-M3 20.6%
  61. 61 GPT-6 Luna 19.9%
  62. 62 Mercury 3 18.2%
  63. 63 Gemini 3.6 Flash 18%
  64. 64 GPT-6.1 Terra 17.1%
  65. 65 Qwen3.7-Max 15.8%
  66. 66 Inkling-Small 15.4%
  67. 67 Inkling 15.3%
  68. 68 GPT-6 Sol 15%
  69. 69 Kimi K2.6 12.4%
  70. 70 Gemini 3 Flash 11.8%
  71. 71 Nemotron 3.5 Super VL 11.6%
  72. 72 MiMo-V2.5-Pro 11.5%
  73. 73 Qwen3.7-Plus 10.5%
  74. 74 Gemini 3.1 Pro 9.29%
  75. 75 GPT-5.4 Mini 8.81%
  76. 76 Beam 1.0 8.81%
  77. 77 MiMo-V2.5 7.77%
  78. 78 Qwen 3.6 Plus 7.68%
  79. 79 Gemini 3.5 Flash-Lite 7.25%
  80. 80 Grok 4.20 reasoning 6.95%
  81. 81 MiniMax-M2.7 6.87%
  82. 82 Grok 4.3 6.18%
  83. 83 Claude Haiku 4.5 thinking 5.86%
  84. 84 GPT-5.4 Nano 4.97%
  85. 85 Mistral Medium 3.5 4.48%
  86. 86 Gemini 3.1 Flash-Lite 3.11%
  87. 87 Nemotron 3.5 Lightning 30B-A3B 2.59%
  88. 88 Mercury 2.5 2.49%

Source: Official Tax Agent Bench leaderboard

87 models. Scroll sideways for the rest.

  1. 52.3 Gemini 4 Hydrogen
  2. 51.8 Gemini 4 Carbon
  3. 50.8 Claude Fable 5.1
  4. 49.7 Qwen3.8 Max (0902)
  5. 49.2 Gemini 4 Argon
  6. 47.1 Claude Opus 5
  7. 44.5 Claude Opus 5.5
  8. 44.4 DeepSeek V4.1 Flash 1017
  9. 43.3 GLM-5.3
  10. 43.0 Muse Wonder Full
  11. 42.4 Muse Spark 1.3
  12. 42.0 Claude Sonnet 5.5
  13. 40.4 Qwen4-Max
  14. 38.9 Inkling 2 Small
  15. 38.3 Claude Fable 5
  16. 38.2 Muse Ember
  17. 36.3 Muse Wonder
  18. 35.8 Grok 4.7
  19. 34.4 Tencent Hy4
  20. 33.7 Grok 4.6
  21. 33.4 Kimi K3
  22. 32.8 Muse Spark 1.2
  23. 32.6 Tencent Hy4 preview
  24. 32.4 MiMo-V2.6 Pro
  25. 32.1 MiniMax-M3.1-Flash
  26. 32.1 Gemini 3.8 Flash
  27. 32.1 Qwen 3.8 Max
  28. 32.1 Claude Fable 5.2
  29. 31.9 Claude Haiku 5.5
  30. 31.6 GLM-5.3-Flash
  31. 31.1 GPT-5.6 Sol
  32. 31.1 Ember-1
  33. 30.9 Qwen 3.8 27B
  34. 30.7 Mistral Large 4
  35. 30.4 Mistral Large 4 Preview
  36. 30.0 GPT-5.6 Terra
  37. 30.0 Claude Opus 4.8
  38. 29.4 Claude Sonnet 5
  39. 28.5 MiMo-V2.6 Flash
  40. 28.1 DeepSeek V4 Flash 0731
  41. 27.3 Step 5 Preview
  42. 26.8 DeepSeek V4 Pro 0813
  43. 26.6 DeepSeek V4.1 Flash
  44. 26.3 Muse Spark Open
  45. 25.8 Claude Sonnet 4.6
  46. 25.2 DeepSeek-V4-Pro
  47. 24.9 Kimi K3.1 Pro
  48. 24.9 Grok 4.5
  49. 24.8 GPT-6.2 Sol
  50. 24.8 GPT-5.6 Luna
  51. 23.7 Claude Opus 4.7
  52. 22.8 Gemini 3.7 Flash
  53. 21.8 GPT-6.2 Astra
  54. 21.8 GPT-5.5
  55. 21.8 GPT-6.1 Sol
  56. 21.6 Gemini 3.5 Flash
  57. 20.7 GPT-6 Astra
  58. 20.6 Ling-3.0-flash-Fin
  59. 20.6 MiniMax-M3
  60. 19.9 GPT-6 Luna
  61. 18.2 Mercury 3
  62. 18.0 Gemini 3.6 Flash
  63. 17.1 GPT-6.1 Terra
  64. 15.8 Qwen3.7-Max
  65. 15.4 Inkling-Small
  66. 15.3 Inkling
  67. 15.0 GPT-6 Sol
  68. 12.4 Kimi K2.6
  69. 11.8 Gemini 3 Flash
  70. 11.6 Nemotron 3.5 Super VL
  71. 11.5 MiMo-V2.5-Pro
  72. 10.5 Qwen3.7-Plus
  73. 9.29 Gemini 3.1 Pro
  74. 8.81 GPT-5.4 Mini
  75. 8.81 Beam 1.0
  76. 7.77 MiMo-V2.5
  77. 7.68 Qwen 3.6 Plus
  78. 7.25 Gemini 3.5 Flash-Lite
  79. 6.95 Grok 4.20
  80. 6.87 MiniMax-M2.7
  81. 6.18 Grok 4.3
  82. 5.86 Claude Haiku 4.5
  83. 4.97 GPT-5.4 Nano
  84. 4.48 Mistral Medium 3.5
  85. 3.11 Gemini 3.1 Flash-Lite
  86. 2.59 Nemotron 3.5 Lightning 30B-A3B
  87. 2.49 Mercury 2.5
  1. 1 Gemini 4 Hydrogen 52.3%
  2. 2 Gemini 4 Carbon 51.8%
  3. 3 Claude Fable 5.1 50.8%
  4. 4 Qwen3.8 Max (0902) 49.7%
  5. 5 Gemini 4 Argon 49.2%
  6. 6 Claude Opus 5 47.1%
  7. 7 Claude Opus 5.5 44.5%
  8. 8 DeepSeek V4.1 Flash 1017 44.4%
  9. 9 GLM-5.3 43.3%
  10. 10 Muse Wonder Full 43%
  11. 11 Muse Spark 1.3 max 42.4%
  12. 12 Claude Sonnet 5.5 42%
  13. 13 Qwen4-Max 40.4%
  14. 14 Inkling 2 Small 38.9%
  15. 15 Claude Fable 5 38.3%
  16. 16 Muse Ember 38.2%
  17. 17 Muse Wonder 36.3%
  18. 18 Grok 4.7 35.8%
  19. 19 Tencent Hy4 34.4%
  20. 20 Grok 4.6 33.7%
  21. 21 Kimi K3 33.4%
  22. 22 Muse Spark 1.2 32.8%
  23. 23 Tencent Hy4 preview 32.6%
  24. 24 MiMo-V2.6 Pro 32.4%
  25. 25 MiniMax-M3.1-Flash 32.1%
  26. 26 Gemini 3.8 Flash 32.1%
  27. 27 Qwen 3.8 Max 32.1%
  28. 28 Claude Fable 5.2 32.1%
  29. 29 Claude Haiku 5.5 31.9%
  30. 30 GLM-5.3-Flash 31.6%
  31. 31 GPT-5.6 Sol 31.1%
  32. 32 Ember-1 31.1%
  33. 33 Qwen 3.8 27B 30.9%
  34. 34 Mistral Large 4 30.7%
  35. 35 Mistral Large 4 Preview 30.4%
  36. 36 GPT-5.6 Terra 30%
  37. 37 Claude Opus 4.8 30%
  38. 38 Claude Sonnet 5 29.4%
  39. 39 MiMo-V2.6 Flash 28.5%
  40. 40 DeepSeek V4 Flash 0731 28.1%
  41. 41 Step 5 Preview 27.3%
  42. 42 DeepSeek V4 Pro 0813 26.8%
  43. 43 DeepSeek V4.1 Flash 26.6%
  44. 44 Muse Spark Open 26.3%
  45. 45 Claude Sonnet 4.6 25.8%
  46. 46 DeepSeek-V4-Pro 25.2%
  47. 47 Kimi K3.1 Pro 24.9%
  48. 48 Grok 4.5 24.9%
  49. 49 GPT-6.2 Sol 24.8%
  50. 50 GPT-5.6 Luna 24.8%
  51. 51 Claude Opus 4.7 23.7%
  52. 52 Gemini 3.7 Flash 22.8%
  53. 53 GPT-6.2 Astra 21.8%
  54. 54 GPT-5.5 21.8%
  55. 55 GPT-6.1 Sol 21.8%
  56. 56 Gemini 3.5 Flash 21.6%
  57. 57 GPT-6 Astra 20.7%
  58. 58 Ling-3.0-flash-Fin 20.6%
  59. 59 MiniMax-M3 20.6%
  60. 60 GPT-6 Luna 19.9%
  61. 61 Mercury 3 18.2%
  62. 62 Gemini 3.6 Flash 18%
  63. 63 GPT-6.1 Terra 17.1%
  64. 64 Qwen3.7-Max 15.8%
  65. 65 Inkling-Small 15.4%
  66. 66 Inkling 15.3%
  67. 67 GPT-6 Sol 15%
  68. 68 Kimi K2.6 12.4%
  69. 69 Gemini 3 Flash 11.8%
  70. 70 Nemotron 3.5 Super VL 11.6%
  71. 71 MiMo-V2.5-Pro 11.5%
  72. 72 Qwen3.7-Plus 10.5%
  73. 73 Gemini 3.1 Pro 9.29%
  74. 74 GPT-5.4 Mini 8.81%
  75. 75 Beam 1.0 8.81%
  76. 76 MiMo-V2.5 7.77%
  77. 77 Qwen 3.6 Plus 7.68%
  78. 78 Gemini 3.5 Flash-Lite 7.25%
  79. 79 Grok 4.20 reasoning 6.95%
  80. 80 MiniMax-M2.7 6.87%
  81. 81 Grok 4.3 6.18%
  82. 82 Claude Haiku 4.5 thinking 5.86%
  83. 83 GPT-5.4 Nano 4.97%
  84. 84 Mistral Medium 3.5 4.48%
  85. 85 Gemini 3.1 Flash-Lite 3.11%
  86. 86 Nemotron 3.5 Lightning 30B-A3B 2.59%
  87. 87 Mercury 2.5 2.49%

Source: Official Tax Agent Bench leaderboard

88 models. Scroll sideways for the rest.

  1. 52.3 Gemini 4 Hydrogen
  2. 51.8 Gemini 4 Carbon
  3. 50.8 Claude Fable 5.1
  4. 49.7 Qwen3.8 Max (0902)
  5. 49.2 Gemini 4 Argon
  6. 47.1 Claude Opus 5
  7. 44.5 Claude Opus 5.5
  8. 44.4 DeepSeek V4.1 Flash 1017
  9. 43.3 GLM-5.3
  10. 43.0 Muse Wonder Full
  11. 42.4 Muse Spark 1.3
  12. 42.0 Claude Sonnet 5.5
  13. 40.4 Qwen4-Max
  14. 38.9 Inkling 2 Small
  15. 38.3 Claude Fable 5
  16. 38.2 Muse Ember
  17. 36.3 Muse Wonder
  18. 35.8 Grok 4.7
  19. 34.4 Tencent Hy4
  20. 33.7 Grok 4.6
  21. 33.4 Kimi K3
  22. 32.8 Muse Spark 1.2
  23. 32.6 Tencent Hy4 preview
  24. 32.4 MiMo-V2.6 Pro
  25. 32.1 MiniMax-M3.1-Flash
  26. 32.1 Gemini 3.8 Flash
  27. 32.1 Qwen 3.8 Max
  28. 32.1 Claude Fable 5.2
  29. 31.9 Claude Haiku 5.5
  30. 31.6 GLM-5.3-Flash
  31. 31.1 GPT-5.6 Sol
  32. 31.1 Ember-1
  33. 30.9 Qwen 3.8 27B
  34. 30.7 Mistral Large 4
  35. 30.4 Mistral Large 4 Preview
  36. 30.0 GPT-5.6 Terra
  37. 30.0 Claude Opus 4.8
  38. 29.4 Claude Sonnet 5
  39. 28.5 MiMo-V2.6 Flash
  40. 28.1 DeepSeek V4 Flash 0731
  41. 27.3 Step 5 Preview
  42. 26.8 DeepSeek V4 Pro 0813
  43. 26.6 DeepSeek V4.1 Flash
  44. 26.3 Muse Spark Open
  45. 25.8 Claude Sonnet 4.6
  46. 25.2 DeepSeek-V4-Pro
  47. 24.9 Kimi K3.1 Pro
  48. 24.9 Grok 4.5
  49. 24.8 GPT-6.2 Sol
  50. 24.8 GPT-5.6 Luna
  51. 23.7 Claude Opus 4.7
  52. 22.8 Gemini 3.7 Flash
  53. 21.8 GPT-6.2 Astra
  54. 21.8 GPT-5.5
  55. 21.8 GPT-6.1 Sol
  56. 21.6 Gemini 3.5 Flash
  57. 20.7 GPT-6 Astra
  58. 20.6 Ling-3.0-flash-Fin
  59. 20.6 MiniMax-M3
  60. 19.9 GPT-6 Luna
  61. 18.2 Mercury 3
  62. 18.0 Gemini 3.6 Flash
  63. 17.1 GPT-6.1 Terra
  64. 15.8 Qwen3.7-Max
  65. 15.4 Inkling-Small
  66. 15.3 Inkling
  67. 15.0 GPT-6 Sol
  68. 14.1 Inkling 2 Mini
  69. 12.4 Kimi K2.6
  70. 11.8 Gemini 3 Flash
  71. 11.6 Nemotron 3.5 Super VL
  72. 11.5 MiMo-V2.5-Pro
  73. 10.5 Qwen3.7-Plus
  74. 9.29 Gemini 3.1 Pro
  75. 8.81 GPT-5.4 Mini
  76. 8.81 Beam 1.0
  77. 7.77 MiMo-V2.5
  78. 7.68 Qwen 3.6 Plus
  79. 7.25 Gemini 3.5 Flash-Lite
  80. 6.95 Grok 4.20
  81. 6.87 MiniMax-M2.7
  82. 6.18 Grok 4.3
  83. 5.86 Claude Haiku 4.5
  84. 4.97 GPT-5.4 Nano
  85. 4.48 Mistral Medium 3.5
  86. 3.11 Gemini 3.1 Flash-Lite
  87. 2.59 Nemotron 3.5 Lightning 30B-A3B
  88. 2.49 Mercury 2.5
  1. 1 Gemini 4 Hydrogen 52.3%
  2. 2 Gemini 4 Carbon 51.8%
  3. 3 Claude Fable 5.1 50.8%
  4. 4 Qwen3.8 Max (0902) 49.7%
  5. 5 Gemini 4 Argon 49.2%
  6. 6 Claude Opus 5 47.1%
  7. 7 Claude Opus 5.5 44.5%
  8. 8 DeepSeek V4.1 Flash 1017 44.4%
  9. 9 GLM-5.3 43.3%
  10. 10 Muse Wonder Full 43%
  11. 11 Muse Spark 1.3 max 42.4%
  12. 12 Claude Sonnet 5.5 42%
  13. 13 Qwen4-Max 40.4%
  14. 14 Inkling 2 Small 38.9%
  15. 15 Claude Fable 5 38.3%
  16. 16 Muse Ember 38.2%
  17. 17 Muse Wonder 36.3%
  18. 18 Grok 4.7 35.8%
  19. 19 Tencent Hy4 34.4%
  20. 20 Grok 4.6 33.7%
  21. 21 Kimi K3 33.4%
  22. 22 Muse Spark 1.2 32.8%
  23. 23 Tencent Hy4 preview 32.6%
  24. 24 MiMo-V2.6 Pro 32.4%
  25. 25 MiniMax-M3.1-Flash 32.1%
  26. 26 Gemini 3.8 Flash 32.1%
  27. 27 Qwen 3.8 Max 32.1%
  28. 28 Claude Fable 5.2 32.1%
  29. 29 Claude Haiku 5.5 31.9%
  30. 30 GLM-5.3-Flash 31.6%
  31. 31 GPT-5.6 Sol 31.1%
  32. 32 Ember-1 31.1%
  33. 33 Qwen 3.8 27B 30.9%
  34. 34 Mistral Large 4 30.7%
  35. 35 Mistral Large 4 Preview 30.4%
  36. 36 GPT-5.6 Terra 30%
  37. 37 Claude Opus 4.8 30%
  38. 38 Claude Sonnet 5 29.4%
  39. 39 MiMo-V2.6 Flash 28.5%
  40. 40 DeepSeek V4 Flash 0731 28.1%
  41. 41 Step 5 Preview 27.3%
  42. 42 DeepSeek V4 Pro 0813 26.8%
  43. 43 DeepSeek V4.1 Flash 26.6%
  44. 44 Muse Spark Open 26.3%
  45. 45 Claude Sonnet 4.6 25.8%
  46. 46 DeepSeek-V4-Pro 25.2%
  47. 47 Kimi K3.1 Pro 24.9%
  48. 48 Grok 4.5 24.9%
  49. 49 GPT-6.2 Sol 24.8%
  50. 50 GPT-5.6 Luna 24.8%
  51. 51 Claude Opus 4.7 23.7%
  52. 52 Gemini 3.7 Flash 22.8%
  53. 53 GPT-6.2 Astra 21.8%
  54. 54 GPT-5.5 21.8%
  55. 55 GPT-6.1 Sol 21.8%
  56. 56 Gemini 3.5 Flash 21.6%
  57. 57 GPT-6 Astra 20.7%
  58. 58 Ling-3.0-flash-Fin 20.6%
  59. 59 MiniMax-M3 20.6%
  60. 60 GPT-6 Luna 19.9%
  61. 61 Mercury 3 18.2%
  62. 62 Gemini 3.6 Flash 18%
  63. 63 GPT-6.1 Terra 17.1%
  64. 64 Qwen3.7-Max 15.8%
  65. 65 Inkling-Small 15.4%
  66. 66 Inkling 15.3%
  67. 67 GPT-6 Sol 15%
  68. 68 Inkling 2 Mini 14.1%
  69. 69 Kimi K2.6 12.4%
  70. 70 Gemini 3 Flash 11.8%
  71. 71 Nemotron 3.5 Super VL 11.6%
  72. 72 MiMo-V2.5-Pro 11.5%
  73. 73 Qwen3.7-Plus 10.5%
  74. 74 Gemini 3.1 Pro 9.29%
  75. 75 GPT-5.4 Mini 8.81%
  76. 76 Beam 1.0 8.81%
  77. 77 MiMo-V2.5 7.77%
  78. 78 Qwen 3.6 Plus 7.68%
  79. 79 Gemini 3.5 Flash-Lite 7.25%
  80. 80 Grok 4.20 reasoning 6.95%
  81. 81 MiniMax-M2.7 6.87%
  82. 82 Grok 4.3 6.18%
  83. 83 Claude Haiku 4.5 thinking 5.86%
  84. 84 GPT-5.4 Nano 4.97%
  85. 85 Mistral Medium 3.5 4.48%
  86. 86 Gemini 3.1 Flash-Lite 3.11%
  87. 87 Nemotron 3.5 Lightning 30B-A3B 2.59%
  88. 88 Mercury 2.5 2.49%

Sources: Official Tax Agent Bench leaderboard and labs’ model cards

#EntryScore ± 95% CISettingDateSource
1 Gemini 4 Hydrogen 52.3% ±7.1 Nov 2026 Official leaderboard
2 Gemini 4 Carbon 51.8% ±7.1 Oct 2026 Official leaderboard
3 Claude Fable 5.1 50.8% ±7.1 Sep 2026 Official leaderboard
4 Qwen3.8 Max (0902) 49.7% ±7.1 Sep 2026 Official leaderboard
5 Gemini 4 Argon 49.2% ±7.1 Sep 2026 Official leaderboard
6 Claude Opus 5 47.1% ±7.1 Jul 2026 Official leaderboard
7 Claude Opus 5.5 44.5% ±7 Sep 2026 Official leaderboard
8 DeepSeek V4.1 Flash 1017 44.4% ±7 Oct 2026 Official leaderboard
9 GLM-5.3 43.3% ±7 Aug 2026 Official leaderboard
10 Muse Wonder Full 43% ±7 Oct 2026 Official leaderboard
11 Muse Spark 1.3 42.4% ±6.9 max Sep 2026 Official leaderboard
12 Claude Sonnet 5.5 42% ±7 Sep 2026 Official leaderboard
13 Qwen4-Max 40.4% ±6.9 Nov 2026 Official leaderboard
14 Inkling 2 Small 38.9% ±6.4 Nov 2026 Official leaderboard
15 Claude Fable 5 38.3% ±6.9 Jun 2026 Official leaderboard
16 Muse Ember 38.2% ±6.9 Nov 2026 Official leaderboard
17 Muse Wonder 36.3% ±6.8 Oct 2026 Official leaderboard
18 Grok 4.7 35.8% ±6.8 Sep 2026 Official leaderboard
19 Tencent Hy4 34.4% ±6.5 Nov 2026 Official leaderboard
20 Grok 4.6 33.7% ±6.7 Aug 2026 Official leaderboard
21 Kimi K3 33.4% ±6.6 Jul 2026 Official leaderboard
22 Muse Spark 1.2 32.8% ±5.4 Aug 2026 Official leaderboard
23 Tencent Hy4 preview 32.6% ±6.6 Aug 2026 Official leaderboard
24 MiMo-V2.6 Pro 32.4% ±6.6 Sep 2026 Official leaderboard
25 MiniMax-M3.1-Flash 32.1% ±6.6 Oct 2026 Official leaderboard
26 Gemini 3.8 Flash 32.1% ±6.6 Sep 2026 Official leaderboard
27 Qwen 3.8 Max 32.1% ±6.6 Aug 2026 Official leaderboard
28 Claude Fable 5.2 32.1% ±6.6 Nov 2026 Official leaderboard
29 Claude Haiku 5.5 31.9% ±6.6 Oct 2026 Official leaderboard
30 GLM-5.3-Flash 31.6% ±6.6 Aug 2026 Official leaderboard
31 GPT-5.6 Sol 31.1% ±6.5 Jul 2026 Official leaderboard
32 Ember-1 31.1% ±6.5 Sep 2026 Official leaderboard
33 Qwen 3.8 27B 30.9% ±6.5 Aug 2026 Official leaderboard
34 Mistral Large 4 30.7% ±6.5 Oct 2026 Official leaderboard
35 Mistral Large 4 Preview 30.4% ±6.5 Oct 2026 Official leaderboard
36 GPT-5.6 Terra 30% ±6.5 Jul 2026 Official leaderboard
37 Claude Opus 4.8 30% ±6.5 May 2026 Official leaderboard
38 Claude Sonnet 5 29.4% ±6.4 Jun 2026 Official leaderboard
39 MiMo-V2.6 Flash 28.5% ±6.2 Sep 2026 Official leaderboard
40 DeepSeek V4 Flash 0731 28.1% ±6.2 Jul 2026 Official leaderboard
41 Step 5 Preview 27.3% ±6.1 Sep 2026 Official leaderboard
42 DeepSeek V4 Pro 0813 26.8% ±6.1 Aug 2026 Official leaderboard
43 DeepSeek V4.1 Flash 26.6% ±6.2 Sep 2026 Official leaderboard
44 Muse Spark Open 26.3% ±5.2 Oct 2026 Official leaderboard
45 Claude Sonnet 4.6 25.8% ±6.2 Feb 2026 Official leaderboard
46 DeepSeek-V4-Pro 25.2% ±6 Apr 2026 Official leaderboard
47 Kimi K3.1 Pro 24.9% ±6.1 Oct 2026 Official leaderboard
48 Grok 4.5 24.9% ±6.1 Jul 2026 Official leaderboard
49 GPT-6.2 Sol 24.8% ±6.1 Nov 2026 Official leaderboard
50 GPT-5.6 Luna 24.8% ±6.1 Jul 2026 Official leaderboard
51 Claude Opus 4.7 23.7% ±6 Apr 2026 Official leaderboard
52 Hy4 Flash 23.3% ±6 Official leaderboard
53 Gemini 3.7 Flash 22.8% ±5.9 Aug 2026 Official leaderboard
54 GPT-6.2 Astra 21.8% ±5.5 Nov 2026 Official leaderboard
55 GPT-5.5 21.8% ±5.8 Apr 2026 Official leaderboard
56 GPT-6.1 Sol 21.8% ±5.8 Sep 2026 Official leaderboard
57 Gemini 3.5 Flash 21.6% ±5.8 May 2026 Official leaderboard
58 GPT-6 Astra 20.7% ±5.7 Sep 2026 Official leaderboard
59 Ling-3.0-flash-Fin 20.6% ±5.7 Sep 2026 Official leaderboard
60 MiniMax-M3 20.6% ±5.6 Jun 2026 Official leaderboard
61 GPT-6 Luna 19.9% ±5.6 Sep 2026 Official leaderboard
62 Mercury 3 18.2% ±5 Oct 2026 Official leaderboard
63 Gemini 3.6 Flash 18% ±5.4 Jul 2026 Official leaderboard
64 GPT-6.1 Terra 17.1% ±5.3 Oct 2026 Official leaderboard
65 Qwen3.7-Max 15.8% ±5.1 May 2026 Official leaderboard
66 Inkling-Small 15.4% ±4.6 Jul 2026 Official leaderboard
67 Inkling 15.3% ±4.7 Jul 2026 Official leaderboard
68 GPT-6 Sol 15% ±5 Sep 2026 Official leaderboard
69 Kimi K2.6 12.4% ±4.7 Apr 2026 Official leaderboard
70 Gemini 3 Flash 11.8% ±4.6 Dec 2025 Official leaderboard
71 Nemotron 3.5 Super VL 11.6% ±4.3 Oct 2026 Official leaderboard
72 MiMo-V2.5-Pro 11.5% ±3.8 Apr 2026 Official leaderboard
73 Qwen3.7-Plus 10.5% ±3.9 Jun 2026 Official leaderboard
74 Gemini 3.1 Pro 9.29% ±4.1 Feb 2026 Official leaderboard
75 GPT-5.4 Mini 8.81% ±4 Mar 2026 Official leaderboard
76 Beam 1.0 8.81% ±3.7 Oct 2026 Official leaderboard
77 MiMo-V2.5 7.77% ±3.5 Apr 2026 Official leaderboard
78 Qwen 3.6 Plus 7.68% ±3.5 Mar 2026 Official leaderboard
79 Gemini 3.5 Flash-Lite 7.25% ±3.7 Jul 2026 Official leaderboard
80 Grok 4.20 6.95% ±3.4 reasoning Feb 2026 Official leaderboard
81 MiniMax-M2.7 6.87% ±3.3 Mar 2026 Official leaderboard
82 Grok 4.3 6.18% ±3.4 Apr 2026 Official leaderboard
83 Claude Haiku 4.5 5.86% ±3 thinking Oct 2025 Official leaderboard
84 GPT-5.4 Nano 4.97% ±2.8 Mar 2026 Official leaderboard
85 Mistral Medium 3.5 4.48% ±2.9 Apr 2026 Official leaderboard
86 Gemini 3.1 Flash-Lite 3.11% ±2.5 Mar 2026 Official leaderboard
87 Nemotron 3.5 Lightning 30B-A3B 2.59% ±2.2 Aug 2026 Official leaderboard
88 Mercury 2.5 2.49% ±2 Sep 2026 Official leaderboard
#ModelScore ± 95% CISettingDateSource
1 Gemini 4 Hydrogen 52.3% ±7.1 Nov 2026 Official leaderboard
2 Gemini 4 Carbon 51.8% ±7.1 Oct 2026 Official leaderboard
3 Claude Fable 5.1 50.8% ±7.1 Sep 2026 Official leaderboard
4 Qwen3.8 Max (0902) 49.7% ±7.1 Sep 2026 Official leaderboard
5 Gemini 4 Argon 49.2% ±7.1 Sep 2026 Official leaderboard
6 Claude Opus 5 47.1% ±7.1 Jul 2026 Official leaderboard
7 Claude Opus 5.5 44.5% ±7 Sep 2026 Official leaderboard
8 DeepSeek V4.1 Flash 1017 44.4% ±7 Oct 2026 Official leaderboard
9 GLM-5.3 43.3% ±7 Aug 2026 Official leaderboard
10 Muse Wonder Full 43% ±7 Oct 2026 Official leaderboard
11 Muse Spark 1.3 42.4% ±6.9 max Sep 2026 Official leaderboard
12 Claude Sonnet 5.5 42% ±7 Sep 2026 Official leaderboard
13 Qwen4-Max 40.4% ±6.9 Nov 2026 Official leaderboard
14 Inkling 2 Small 38.9% ±6.4 Nov 2026 Official leaderboard
15 Claude Fable 5 38.3% ±6.9 Jun 2026 Official leaderboard
16 Muse Ember 38.2% ±6.9 Nov 2026 Official leaderboard
17 Muse Wonder 36.3% ±6.8 Oct 2026 Official leaderboard
18 Grok 4.7 35.8% ±6.8 Sep 2026 Official leaderboard
19 Tencent Hy4 34.4% ±6.5 Nov 2026 Official leaderboard
20 Grok 4.6 33.7% ±6.7 Aug 2026 Official leaderboard
21 Kimi K3 33.4% ±6.6 Jul 2026 Official leaderboard
22 Muse Spark 1.2 32.8% ±5.4 Aug 2026 Official leaderboard
23 Tencent Hy4 preview 32.6% ±6.6 Aug 2026 Official leaderboard
24 MiMo-V2.6 Pro 32.4% ±6.6 Sep 2026 Official leaderboard
25 MiniMax-M3.1-Flash 32.1% ±6.6 Oct 2026 Official leaderboard
26 Gemini 3.8 Flash 32.1% ±6.6 Sep 2026 Official leaderboard
27 Qwen 3.8 Max 32.1% ±6.6 Aug 2026 Official leaderboard
28 Claude Fable 5.2 32.1% ±6.6 Nov 2026 Official leaderboard
29 Claude Haiku 5.5 31.9% ±6.6 Oct 2026 Official leaderboard
30 GLM-5.3-Flash 31.6% ±6.6 Aug 2026 Official leaderboard
31 GPT-5.6 Sol 31.1% ±6.5 Jul 2026 Official leaderboard
32 Ember-1 31.1% ±6.5 Sep 2026 Official leaderboard
33 Qwen 3.8 27B 30.9% ±6.5 Aug 2026 Official leaderboard
34 Mistral Large 4 30.7% ±6.5 Oct 2026 Official leaderboard
35 Mistral Large 4 Preview 30.4% ±6.5 Oct 2026 Official leaderboard
36 GPT-5.6 Terra 30% ±6.5 Jul 2026 Official leaderboard
37 Claude Opus 4.8 30% ±6.5 May 2026 Official leaderboard
38 Claude Sonnet 5 29.4% ±6.4 Jun 2026 Official leaderboard
39 MiMo-V2.6 Flash 28.5% ±6.2 Sep 2026 Official leaderboard
40 DeepSeek V4 Flash 0731 28.1% ±6.2 Jul 2026 Official leaderboard
41 Step 5 Preview 27.3% ±6.1 Sep 2026 Official leaderboard
42 DeepSeek V4 Pro 0813 26.8% ±6.1 Aug 2026 Official leaderboard
43 DeepSeek V4.1 Flash 26.6% ±6.2 Sep 2026 Official leaderboard
44 Muse Spark Open 26.3% ±5.2 Oct 2026 Official leaderboard
45 Claude Sonnet 4.6 25.8% ±6.2 Feb 2026 Official leaderboard
46 DeepSeek-V4-Pro 25.2% ±6 Apr 2026 Official leaderboard
47 Kimi K3.1 Pro 24.9% ±6.1 Oct 2026 Official leaderboard
48 Grok 4.5 24.9% ±6.1 Jul 2026 Official leaderboard
49 GPT-6.2 Sol 24.8% ±6.1 Nov 2026 Official leaderboard
50 GPT-5.6 Luna 24.8% ±6.1 Jul 2026 Official leaderboard
51 Claude Opus 4.7 23.7% ±6 Apr 2026 Official leaderboard
52 Gemini 3.7 Flash 22.8% ±5.9 Aug 2026 Official leaderboard
53 GPT-6.2 Astra 21.8% ±5.5 Nov 2026 Official leaderboard
54 GPT-5.5 21.8% ±5.8 Apr 2026 Official leaderboard
55 GPT-6.1 Sol 21.8% ±5.8 Sep 2026 Official leaderboard
56 Gemini 3.5 Flash 21.6% ±5.8 May 2026 Official leaderboard
57 GPT-6 Astra 20.7% ±5.7 Sep 2026 Official leaderboard
58 Ling-3.0-flash-Fin 20.6% ±5.7 Sep 2026 Official leaderboard
59 MiniMax-M3 20.6% ±5.6 Jun 2026 Official leaderboard
60 GPT-6 Luna 19.9% ±5.6 Sep 2026 Official leaderboard
61 Mercury 3 18.2% ±5 Oct 2026 Official leaderboard
62 Gemini 3.6 Flash 18% ±5.4 Jul 2026 Official leaderboard
63 GPT-6.1 Terra 17.1% ±5.3 Oct 2026 Official leaderboard
64 Qwen3.7-Max 15.8% ±5.1 May 2026 Official leaderboard
65 Inkling-Small 15.4% ±4.6 Jul 2026 Official leaderboard
66 Inkling 15.3% ±4.7 Jul 2026 Official leaderboard
67 GPT-6 Sol 15% ±5 Sep 2026 Official leaderboard
68 Kimi K2.6 12.4% ±4.7 Apr 2026 Official leaderboard
69 Gemini 3 Flash 11.8% ±4.6 Dec 2025 Official leaderboard
70 Nemotron 3.5 Super VL 11.6% ±4.3 Oct 2026 Official leaderboard
71 MiMo-V2.5-Pro 11.5% ±3.8 Apr 2026 Official leaderboard
72 Qwen3.7-Plus 10.5% ±3.9 Jun 2026 Official leaderboard
73 Gemini 3.1 Pro 9.29% ±4.1 Feb 2026 Official leaderboard
74 GPT-5.4 Mini 8.81% ±4 Mar 2026 Official leaderboard
75 Beam 1.0 8.81% ±3.7 Oct 2026 Official leaderboard
76 MiMo-V2.5 7.77% ±3.5 Apr 2026 Official leaderboard
77 Qwen 3.6 Plus 7.68% ±3.5 Mar 2026 Official leaderboard
78 Gemini 3.5 Flash-Lite 7.25% ±3.7 Jul 2026 Official leaderboard
79 Grok 4.20 6.95% ±3.4 reasoning Feb 2026 Official leaderboard
80 MiniMax-M2.7 6.87% ±3.3 Mar 2026 Official leaderboard
81 Grok 4.3 6.18% ±3.4 Apr 2026 Official leaderboard
82 Claude Haiku 4.5 5.86% ±3 thinking Oct 2025 Official leaderboard
83 GPT-5.4 Nano 4.97% ±2.8 Mar 2026 Official leaderboard
84 Mistral Medium 3.5 4.48% ±2.9 Apr 2026 Official leaderboard
85 Gemini 3.1 Flash-Lite 3.11% ±2.5 Mar 2026 Official leaderboard
86 Nemotron 3.5 Lightning 30B-A3B 2.59% ±2.2 Aug 2026 Official leaderboard
87 Mercury 2.5 2.49% ±2 Sep 2026 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 4 Hydrogen 52.3% Nov 2026 Official leaderboard
2 Gemini 4 Carbon 51.8% Oct 2026 Official leaderboard
3 Claude Fable 5.1 50.8% Sep 2026 Official leaderboard
4 Qwen3.8 Max (0902) 49.7% Sep 2026 Official leaderboard
5 Gemini 4 Argon 49.2% Sep 2026 Official leaderboard
6 Claude Opus 5 47.1% Jul 2026 Official leaderboard
7 Claude Opus 5.5 44.5% Sep 2026 Official leaderboard
8 DeepSeek V4.1 Flash 1017 44.4% Oct 2026 Official leaderboard
9 GLM-5.3 43.3% Aug 2026 Official leaderboard
10 Muse Wonder Full 43% Oct 2026 Official leaderboard
11 Muse Spark 1.3 42.4% max Sep 2026 Official leaderboard
12 Claude Sonnet 5.5 42% Sep 2026 Official leaderboard
13 Qwen4-Max 40.4% Nov 2026 Official leaderboard
14 Inkling 2 Small 38.9% Nov 2026 Official leaderboard
15 Claude Fable 5 38.3% Jun 2026 Official leaderboard
16 Muse Ember 38.2% Nov 2026 Official leaderboard
17 Muse Wonder 36.3% Oct 2026 Official leaderboard
18 Grok 4.7 35.8% Sep 2026 Official leaderboard
19 Tencent Hy4 34.4% Nov 2026 Official leaderboard
20 Grok 4.6 33.7% Aug 2026 Official leaderboard
21 Kimi K3 33.4% Jul 2026 Official leaderboard
22 Muse Spark 1.2 32.8% Aug 2026 Official leaderboard
23 Tencent Hy4 preview 32.6% Aug 2026 Official leaderboard
24 MiMo-V2.6 Pro 32.4% Sep 2026 Official leaderboard
25 MiniMax-M3.1-Flash 32.1% Oct 2026 Official leaderboard
26 Gemini 3.8 Flash 32.1% Sep 2026 Official leaderboard
27 Qwen 3.8 Max 32.1% Aug 2026 Official leaderboard
28 Claude Fable 5.2 32.1% Nov 2026 Official leaderboard
29 Claude Haiku 5.5 31.9% Oct 2026 Official leaderboard
30 GLM-5.3-Flash 31.6% Aug 2026 Official leaderboard
31 GPT-5.6 Sol 31.1% Jul 2026 Official leaderboard
32 Ember-1 31.1% Sep 2026 Official leaderboard
33 Qwen 3.8 27B 30.9% Aug 2026 Official leaderboard
34 Mistral Large 4 30.7% Oct 2026 Official leaderboard
35 Mistral Large 4 Preview 30.4% Oct 2026 Official leaderboard
36 GPT-5.6 Terra 30% Jul 2026 Official leaderboard
37 Claude Opus 4.8 30% May 2026 Official leaderboard
38 Claude Sonnet 5 29.4% Jun 2026 Official leaderboard
39 MiMo-V2.6 Flash 28.5% Sep 2026 Official leaderboard
40 DeepSeek V4 Flash 0731 28.1% Jul 2026 Official leaderboard
41 Step 5 Preview 27.3% Sep 2026 Official leaderboard
42 DeepSeek V4 Pro 0813 26.8% Aug 2026 Official leaderboard
43 DeepSeek V4.1 Flash 26.6% Sep 2026 Official leaderboard
44 Muse Spark Open 26.3% Oct 2026 Official leaderboard
45 Claude Sonnet 4.6 25.8% Feb 2026 Official leaderboard
46 DeepSeek-V4-Pro 25.2% Apr 2026 Official leaderboard
47 Kimi K3.1 Pro 24.9% Oct 2026 Official leaderboard
48 Grok 4.5 24.9% Jul 2026 Official leaderboard
49 GPT-6.2 Sol 24.8% Nov 2026 Official leaderboard
50 GPT-5.6 Luna 24.8% Jul 2026 Official leaderboard
51 Claude Opus 4.7 23.7% Apr 2026 Official leaderboard
52 Gemini 3.7 Flash 22.8% Aug 2026 Official leaderboard
53 GPT-6.2 Astra 21.8% Nov 2026 Official leaderboard
54 GPT-5.5 21.8% Apr 2026 Official leaderboard
55 GPT-6.1 Sol 21.8% Sep 2026 Official leaderboard
56 Gemini 3.5 Flash 21.6% May 2026 Official leaderboard
57 GPT-6 Astra 20.7% Sep 2026 Official leaderboard
58 Ling-3.0-flash-Fin 20.6% Sep 2026 Official leaderboard
59 MiniMax-M3 20.6% Jun 2026 Official leaderboard
60 GPT-6 Luna 19.9% Sep 2026 Official leaderboard
61 Mercury 3 18.2% Oct 2026 Official leaderboard
62 Gemini 3.6 Flash 18% Jul 2026 Official leaderboard
63 GPT-6.1 Terra 17.1% Oct 2026 Official leaderboard
64 Qwen3.7-Max 15.8% May 2026 Official leaderboard
65 Inkling-Small 15.4% Jul 2026 Official leaderboard
66 Inkling 15.3% Jul 2026 Official leaderboard
67 GPT-6 Sol 15% Sep 2026 Official leaderboard
68 Inkling 2 Mini 14.1% Nov 2026 Lab model card
69 Kimi K2.6 12.4% Apr 2026 Official leaderboard
70 Gemini 3 Flash 11.8% Dec 2025 Official leaderboard
71 Nemotron 3.5 Super VL 11.6% Oct 2026 Official leaderboard
72 MiMo-V2.5-Pro 11.5% Apr 2026 Official leaderboard
73 Qwen3.7-Plus 10.5% Jun 2026 Official leaderboard
74 Gemini 3.1 Pro 9.29% Feb 2026 Official leaderboard
75 GPT-5.4 Mini 8.81% Mar 2026 Official leaderboard
76 Beam 1.0 8.81% Oct 2026 Official leaderboard
77 MiMo-V2.5 7.77% Apr 2026 Official leaderboard
78 Qwen 3.6 Plus 7.68% Mar 2026 Official leaderboard
79 Gemini 3.5 Flash-Lite 7.25% Jul 2026 Official leaderboard
80 Grok 4.20 6.95% reasoning Feb 2026 Official leaderboard
81 MiniMax-M2.7 6.87% Mar 2026 Official leaderboard
82 Grok 4.3 6.18% Apr 2026 Official leaderboard
83 Claude Haiku 4.5 5.86% thinking Oct 2025 Official leaderboard
84 GPT-5.4 Nano 4.97% Mar 2026 Official leaderboard
85 Mistral Medium 3.5 4.48% Apr 2026 Official leaderboard
86 Gemini 3.1 Flash-Lite 3.11% Mar 2026 Official leaderboard
87 Nemotron 3.5 Lightning 30B-A3B 2.59% Aug 2026 Official leaderboard
88 Mercury 2.5 2.49% Sep 2026 Official leaderboard

Labs report scores for 2 models in their cards. Switch the scores above to rank them together.

Other ways to score it

The benchmark’s own leaderboard also reports Weighted pass, another way to score the whole test. Labs sometimes quote it instead of the headline. Gemini 4 Argon leads it. Arrows show how far a model moved from its overall rank among the models we track.

Partial credit: the weighted pass rate the board shows beside the all-pass accuracy. 87 models we track have a score here.

  1. 1 Gemini 4 Argon ↑4 79%
  2. 2 Gemini 4 Hydrogen ↓1 78.4%
  3. 3 Claude Fable 5.1 77.9%
  4. 4 Claude Opus 5 ↑2 76.7%
  5. 5 Gemini 4 Carbon ↓3 76.4%
  6. 6 GLM-5.3 ↑3 75.2%
  7. 7 Muse Wonder Full ↑3 75%
  8. 8 DeepSeek V4.1 Flash 1017 74.9%
  9. 9 Qwen3.8 Max (0902) ↓5 74.3%
  10. 10 Claude Sonnet 5.5 ↑2 73.4%

Accuracy against latency

The Vals scatter chart with its x axis switched from Cost to Latency: mean time per test against accuracy, on a log scale. The line joins the runs that no other run beats on both measures. We redrew this chart from the numbers on the official leaderboard.

0% 20% 40% 60% 101001,00010K Latency (mean per test) (s), log scale, lower is better Accuracy Gemini 4 HydrogenGemini 4 ArgonMuse Spark 1.3 (max)Muse EmberGemini 3.7 FlashNemotron 3.5 Super VLBeam 1.0Gemini 3.5 Flash-LiteGemini 3.1 Flash-Lite 0% 20% 40% 60% 101001,00010K Latency (mean per test) (s), log scale, lower is better Accuracy Gemini 4 HydrogenGemini 4 ArgonMuse Spark 1.3 (max)Muse EmberGemini 3.7 FlashNemotron 3.5 Super VLBeam 1.0Gemini 3.5 Flash-LiteGemini 3.1 Flash-Lite
ModelSettingAccuracyLatency (mean per test)
Gemini 4 Hydrogen 52.3% 1,095 s
Gemini 4 Carbon 51.8% 2,086 s
Claude Fable 5.1 50.8% 1,714 s
Qwen3.8 Max (0902) 49.7% 5,231 s
Gemini 4 Argon 49.2% 985.63 s
Claude Opus 5 47.1% 1,008 s
Claude Opus 5.5 44.5% 4,013 s
DeepSeek V4.1 Flash 1017 44.4% 1,264 s
GLM-5.3 43.3% 3,053 s
Muse Wonder Full 43% 1,084 s
Muse Spark 1.3 max 42.4% 248.82 s
Claude Sonnet 5.5 42% 3,561 s
Qwen4-Max 40.4% 1,473 s
Inkling 2 Small 38.9% 414.04 s
Claude Fable 5 38.3% 866.63 s
Muse Ember 38.2% 90.13 s
Muse Spark 1.3 37.9% 505.37 s
Muse Wonder 36.3% 492.84 s
Grok 4.7 35.8% 996.47 s
Tencent Hy4 34.4% 3,808 s
Grok 4.6 33.7% 657.84 s
Kimi K3 33.4% 2,415 s
Muse Spark 1.2 32.8% 119.91 s
Tencent Hy4 preview 32.6% 2,464 s
MiMo-V2.6 Pro 32.4% 1,548 s
MiniMax-M3.1-Flash 32.1% 1,963 s
Gemini 3.8 Flash 32.1% 178.33 s
Claude Fable 5.2 32.1% 299.74 s
Qwen 3.8 Max 32.1% 2,436 s
Claude Haiku 5.5 31.9% 2,974 s
GLM-5.3-Flash 31.6% 1,297 s
GPT-5.6 Sol 31.1% 2,681 s
Ember-1 31.1% 1,846 s
Qwen 3.8 27B 30.9% 1,334 s
Mistral Large 4 30.7% 1,023 s
Mistral Large 4 Preview 30.4% 3,438 s
GPT-5.6 Terra 30% 3,631 s
Claude Opus 4.8 30% 479.2 s
Claude Sonnet 5 29.4% 1,159 s
MiMo-V2.6 Flash 28.5% 1,039 s
DeepSeek V4 Flash 0731 28.1% 337.34 s
Step 5 Preview 27.3% 1,289 s
DeepSeek V4 Pro 0813 26.8% 1,597 s
DeepSeek V4.1 Flash 26.6% 349.08 s
Muse Spark Open 26.3% 701.83 s
Claude Sonnet 4.6 25.8% 707.31 s
DeepSeek-V4-Pro 25.2% 1,245 s
Kimi K3.1 Pro 24.9% 3,519 s
Grok 4.5 24.9% 283.57 s
GPT-6.2 Sol 24.8% 2,617 s
GPT-5.6 Luna 24.8% 2,566 s
Claude Opus 4.7 23.7% 279.04 s
Hy4 Flash 23.3% 371.72 s
Gemini 3.7 Flash 22.8% 87.24 s
GPT-6.2 Astra 21.8% 3,693 s
GPT-5.5 21.8% 951.81 s
GPT-6.1 Sol 21.8% 1,214 s
Gemini 3.5 Flash 21.7% 233 s
GPT-6 Astra 20.7% 900.59 s
Ling-3.0-flash-Fin 20.6% 333.83 s
MiniMax-M3 20.6% 314.57 s
GPT-6 Luna 20% 1,996 s
Mercury 3 18.2% 269.27 s
Gemini 3.6 Flash 18% 105.32 s
GPT-6.1 Terra 17.1% 2,751 s
Qwen3.7-Max 15.8% 209.73 s
Inkling-Small 15.4% 385.25 s
Inkling 15.3% 810.56 s
GPT-6 Sol 15% 1,052 s
Kimi K2.6 12.4% 924.42 s
Gemini 3 Flash 11.8% 100.62 s
Nemotron 3.5 Super VL 11.6% 72.66 s
MiMo-V2.5-Pro 11.5% 376.51 s
Qwen3.7-Plus 10.5% 310.26 s
Gemini 3.1 Pro 9.3% 100.96 s
Beam 1.0 8.8% 57.99 s
GPT-5.4 Mini 8.8% 1,006 s
MiMo-V2.5 7.8% 500.83 s
Qwen 3.6 Plus 7.7% 378.16 s
Gemini 3.5 Flash-Lite 7.3% 47.46 s
Grok 4.20 reasoning 7% 89.26 s
MiniMax-M2.7 6.9% 302.85 s
Grok 4.3 6.2% 97.52 s
Claude Haiku 4.5 thinking 5.9% 155.02 s
GPT-5.4 Nano 5% 220.31 s
Mistral Medium 3.5 4.5% 430.67 s
Gemini 3.1 Flash-Lite 3.1% 18.74 s
Nemotron 3.5 Lightning 30B-A3B 2.6% 1,180 s
Mercury 2.5 2.5% 29.67 s

Accuracy against latency

The Vals scatter chart with its x axis switched from Cost to Latency: mean time per test against accuracy, on a log scale. The line joins the runs that no other run beats on both measures. We redrew this chart from the numbers on the official leaderboard.

0% 20% 40% 60% 101001,00010K Latency (mean per test) (s), log scale, lower is better Accuracy Gemini 4 HydrogenGemini 4 ArgonMuse Spark 1.3 (max)Muse EmberGemini 3.7 FlashNemotron 3.5 Super VLBeam 1.0Gemini 3.5 Flash-LiteGemini 3.1 Flash-Lite 0% 20% 40% 60% 101001,00010K Latency (mean per test) (s), log scale, lower is better Accuracy Gemini 4 HydrogenGemini 4 ArgonMuse Spark 1.3 (max)Muse EmberGemini 3.7 FlashNemotron 3.5 Super VLBeam 1.0Gemini 3.5 Flash-LiteGemini 3.1 Flash-Lite
ModelSettingAccuracyLatency (mean per test)
Gemini 4 Hydrogen 52.3% 1,095 s
Gemini 4 Carbon 51.8% 2,086 s
Claude Fable 5.1 50.8% 1,714 s
Qwen3.8 Max (0902) 49.7% 5,231 s
Gemini 4 Argon 49.2% 985.63 s
Claude Opus 5 47.1% 1,008 s
Claude Opus 5.5 44.5% 4,013 s
DeepSeek V4.1 Flash 1017 44.4% 1,264 s
GLM-5.3 43.3% 3,053 s
Muse Wonder Full 43% 1,084 s
Muse Spark 1.3 max 42.4% 248.82 s
Claude Sonnet 5.5 42% 3,561 s
Qwen4-Max 40.4% 1,473 s
Inkling 2 Small 38.9% 414.04 s
Claude Fable 5 38.3% 866.63 s
Muse Ember 38.2% 90.13 s
Muse Spark 1.3 37.9% 505.37 s
Muse Wonder 36.3% 492.84 s
Grok 4.7 35.8% 996.47 s
Tencent Hy4 34.4% 3,808 s
Grok 4.6 33.7% 657.84 s
Kimi K3 33.4% 2,415 s
Muse Spark 1.2 32.8% 119.91 s
Tencent Hy4 preview 32.6% 2,464 s
MiMo-V2.6 Pro 32.4% 1,548 s
MiniMax-M3.1-Flash 32.1% 1,963 s
Gemini 3.8 Flash 32.1% 178.33 s
Claude Fable 5.2 32.1% 299.74 s
Qwen 3.8 Max 32.1% 2,436 s
Claude Haiku 5.5 31.9% 2,974 s
GLM-5.3-Flash 31.6% 1,297 s
GPT-5.6 Sol 31.1% 2,681 s
Ember-1 31.1% 1,846 s
Qwen 3.8 27B 30.9% 1,334 s
Mistral Large 4 30.7% 1,023 s
Mistral Large 4 Preview 30.4% 3,438 s
GPT-5.6 Terra 30% 3,631 s
Claude Opus 4.8 30% 479.2 s
Claude Sonnet 5 29.4% 1,159 s
MiMo-V2.6 Flash 28.5% 1,039 s
DeepSeek V4 Flash 0731 28.1% 337.34 s
Step 5 Preview 27.3% 1,289 s
DeepSeek V4 Pro 0813 26.8% 1,597 s
DeepSeek V4.1 Flash 26.6% 349.08 s
Muse Spark Open 26.3% 701.83 s
Claude Sonnet 4.6 25.8% 707.31 s
DeepSeek-V4-Pro 25.2% 1,245 s
Kimi K3.1 Pro 24.9% 3,519 s
Grok 4.5 24.9% 283.57 s
GPT-6.2 Sol 24.8% 2,617 s
GPT-5.6 Luna 24.8% 2,566 s
Claude Opus 4.7 23.7% 279.04 s
Gemini 3.7 Flash 22.8% 87.24 s
GPT-6.2 Astra 21.8% 3,693 s
GPT-5.5 21.8% 951.81 s
GPT-6.1 Sol 21.8% 1,214 s
Gemini 3.5 Flash 21.7% 233 s
GPT-6 Astra 20.7% 900.59 s
Ling-3.0-flash-Fin 20.6% 333.83 s
MiniMax-M3 20.6% 314.57 s
GPT-6 Luna 20% 1,996 s
Mercury 3 18.2% 269.27 s
Gemini 3.6 Flash 18% 105.32 s
GPT-6.1 Terra 17.1% 2,751 s
Qwen3.7-Max 15.8% 209.73 s
Inkling-Small 15.4% 385.25 s
Inkling 15.3% 810.56 s
GPT-6 Sol 15% 1,052 s
Kimi K2.6 12.4% 924.42 s
Gemini 3 Flash 11.8% 100.62 s
Nemotron 3.5 Super VL 11.6% 72.66 s
MiMo-V2.5-Pro 11.5% 376.51 s
Qwen3.7-Plus 10.5% 310.26 s
Gemini 3.1 Pro 9.3% 100.96 s
Beam 1.0 8.8% 57.99 s
GPT-5.4 Mini 8.8% 1,006 s
MiMo-V2.5 7.8% 500.83 s
Qwen 3.6 Plus 7.7% 378.16 s
Gemini 3.5 Flash-Lite 7.3% 47.46 s
Grok 4.20 reasoning 7% 89.26 s
MiniMax-M2.7 6.9% 302.85 s
Grok 4.3 6.2% 97.52 s
Claude Haiku 4.5 thinking 5.9% 155.02 s
GPT-5.4 Nano 5% 220.31 s
Mistral Medium 3.5 4.5% 430.67 s
Gemini 3.1 Flash-Lite 3.1% 18.74 s
Nemotron 3.5 Lightning 30B-A3B 2.6% 1,180 s
Mercury 2.5 2.5% 29.67 s

The best score, 52.3%, costs $3.87 a task

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Google DeepMind Xiaomi Corp Meta DeepSeek Thinking Machines Cost frontier One model at each effort level
0% 20% 40% 60% $0.01 $0.10 $1.00 $10 $100 Cost per task, log scale Gemini 4 Hydrogen Gemini 4 Argon DeepSeek V4.1 Flash 1017 Inkling 2 Small Muse Ember MiMo-V2.5 Gemini 3.1 Flash-Lite 0% 20% 40% 60% $0.01 $0.10 $1.00 $10 $100 Cost per task, log scale Gemini 4 Hydrogen DeepSeek V4.1 Flash 1017

Source: Official Tax Agent Bench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemini 4 Hydrogen 52.3% $3.87 307,129
Gemini 4 Carbon 51.8% $6.34 182,644
Claude Fable 5.1 50.8% $13 146,309
Qwen3.8 Max (0902) 49.7% $4.07 287,749
Gemini 4 Argon 49.2% $3.47 75,098
Claude Opus 5 47.1% $5.14 86,847
Claude Opus 5.5 44.5% $14 463,161
DeepSeek V4.1 Flash 1017 44.4% $0.46 277,522
GLM-5.3 43.3% $1.80 245,685
Muse Wonder Full 43% $2.15 51,650
Muse Spark 1.3 max 42.4% $0.39 31,585
Claude Sonnet 5.5 42% $8.73 565,600
Qwen4-Max 40.4% $1.03 97,273
Inkling 2 Small 38.9% $0.14 72,443
Claude Fable 5 38.3% $6.70 73,091
Muse Ember 38.2% $0.03 32,504
Muse Spark 1.3 37.9% $0.32 24,548
Muse Wonder 36.3% $1.88 45,444
Grok 4.7 35.8% $1.80 56,070
Tencent Hy4 34.4% $1.69 416,480
Grok 4.6 33.7% $0.98 39,758
Kimi K3 33.4% $2.79 79,342
Muse Spark 1.2 32.8% $0.27 19,763
Tencent Hy4 preview 32.6% $0.59 149,760
MiMo-V2.6 Pro 32.4% $0.13 64,880
MiniMax-M3.1-Flash 32.1% $0.32 364,312
Gemini 3.8 Flash 32.1% $0.86 36,335
Qwen 3.8 Max 32.1% $1.55 107,611
Claude Fable 5.2 32.1% $3.31 33,788
Claude Haiku 5.5 31.9% $1.16 741,776
GLM-5.3-Flash 31.6% $0.14 148,072
GPT-5.6 Sol 31.1% $7.51 193,485
Ember-1 31.1% $2.97 71,431
Qwen 3.8 27B 30.9% $0.82 187,156
Mistral Large 4 30.7% $0.50 76,008
Mistral Large 4 Preview 30.4% $1.39 139,787
GPT-5.6 Terra 30% $4.83 268,407
Claude Opus 4.8 30% $1.69 38,739
Claude Sonnet 5 29.4% $1.79 113,409
MiMo-V2.6 Flash 28.5% $0.05 63,189
DeepSeek V4 Flash 0731 28.1% $0.16 76,525
Step 5 Preview 27.3% $0.55 84,980
DeepSeek V4 Pro 0813 26.8% $0.73 109,118
DeepSeek V4.1 Flash 26.6% $0.13 80,656
Muse Spark Open 26.3% $0.88 18,601
Claude Sonnet 4.6 25.8% $1.19 38,465
DeepSeek-V4-Pro 25.2% $0.72 110,422
Kimi K3.1 Pro 24.9% $3.64 180,052
Grok 4.5 24.9% $0.47 16,998
GPT-6.2 Sol 24.8% $2.40 120,431
GPT-5.6 Luna 24.8% $0.47 271,107
Claude Opus 4.7 23.7% $1.27 21,092
Gemini 3.7 Flash 22.8% $0.48 20,337
GPT-6.2 Astra 21.8% $7.90 109,872
GPT-5.5 21.8% $4.07 66,972
GPT-6.1 Sol 21.8% $1.32 65,893
Gemini 3.5 Flash 21.6% $0.86 30,433
GPT-6 Astra 20.7% $5.80 62,346
Ling-3.0-flash-Fin 20.6% $0.03 43,263
MiniMax-M3 20.6% $0.16 24,976
GPT-6 Luna 19.9% $0.19 214,022
Mercury 3 18.2% $0.05 73,519
Gemini 3.6 Flash 18% $0.36 16,362
GPT-6.1 Terra 17.1% $2.16 239,078
Qwen3.7-Max 15.8% $0.44 18,741
Inkling-Small 15.4% $0.09 38,499
Inkling 15.3% $0.35 37,668
GPT-6 Sol 15% $2.28 97,209
Kimi K2.6 12.4% $0.57 45,053
Gemini 3 Flash 11.8% $0.10 16,306
Nemotron 3.5 Super VL 11.6% $0.05 30,297
MiMo-V2.5-Pro 11.5% $0.04 12,625
Qwen3.7-Plus 10.5% $0.09 22,236
Gemini 3.1 Pro 9.29% $0.30 9,861
GPT-5.4 Mini 8.81% $0.81 137,206
Beam 1.0 8.81% 20,848
MiMo-V2.5 7.77% $0.02 15,557
Qwen 3.6 Plus 7.68% $0.30 17,531
Gemini 3.5 Flash-Lite 7.25% $0.08 9,682
Grok 4.20 reasoning 6.95% $0.12 8,326
MiniMax-M2.7 6.87% $0.12 14,118
Grok 4.3 6.18% $0.12 9,232
Claude Haiku 4.5 thinking 5.86% $0.23 14,371
GPT-5.4 Nano 4.97% $0.07 23,086
Mistral Medium 3.5 4.48% $1.44 36,331
Gemini 3.1 Flash-Lite 3.11% $0.01 4,699
Nemotron 3.5 Lightning 30B-A3B 2.59% $0.03 116,422
Mercury 2.5 2.49% $0.02 8,007

The best score, 52.3%, costs $3.87 a task

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Google DeepMind Xiaomi Corp Meta DeepSeek Thinking Machines Cost frontier One model at each effort level
0% 20% 40% 60% $0.01 $0.10 $1.00 $10 $100 Cost per task, log scale Gemini 4 Hydrogen Gemini 4 Argon DeepSeek V4.1 Flash 1017 Inkling 2 Small Muse Ember MiMo-V2.5 Gemini 3.1 Flash-Lite 0% 20% 40% 60% $0.01 $0.10 $1.00 $10 $100 Cost per task, log scale Gemini 4 Hydrogen DeepSeek V4.1 Flash 1017

Source: Official Tax Agent Bench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemini 4 Hydrogen 52.3% $3.87 307,129
Gemini 4 Carbon 51.8% $6.34 182,644
Claude Fable 5.1 50.8% $13 146,309
Qwen3.8 Max (0902) 49.7% $4.07 287,749
Gemini 4 Argon 49.2% $3.47 75,098
Claude Opus 5 47.1% $5.14 86,847
Claude Opus 5.5 44.5% $14 463,161
DeepSeek V4.1 Flash 1017 44.4% $0.46 277,522
GLM-5.3 43.3% $1.80 245,685
Muse Wonder Full 43% $2.15 51,650
Muse Spark 1.3 max 42.4% $0.39 31,585
Claude Sonnet 5.5 42% $8.73 565,600
Qwen4-Max 40.4% $1.03 97,273
Inkling 2 Small 38.9% $0.14 72,443
Claude Fable 5 38.3% $6.70 73,091
Muse Ember 38.2% $0.03 32,504
Muse Spark 1.3 37.9% $0.32 24,548
Muse Wonder 36.3% $1.88 45,444
Grok 4.7 35.8% $1.80 56,070
Tencent Hy4 34.4% $1.69 416,480
Grok 4.6 33.7% $0.98 39,758
Kimi K3 33.4% $2.79 79,342
Muse Spark 1.2 32.8% $0.27 19,763
Tencent Hy4 preview 32.6% $0.59 149,760
MiMo-V2.6 Pro 32.4% $0.13 64,880
MiniMax-M3.1-Flash 32.1% $0.32 364,312
Gemini 3.8 Flash 32.1% $0.86 36,335
Qwen 3.8 Max 32.1% $1.55 107,611
Claude Fable 5.2 32.1% $3.31 33,788
Claude Haiku 5.5 31.9% $1.16 741,776
GLM-5.3-Flash 31.6% $0.14 148,072
GPT-5.6 Sol 31.1% $7.51 193,485
Ember-1 31.1% $2.97 71,431
Qwen 3.8 27B 30.9% $0.82 187,156
Mistral Large 4 30.7% $0.50 76,008
Mistral Large 4 Preview 30.4% $1.39 139,787
GPT-5.6 Terra 30% $4.83 268,407
Claude Opus 4.8 30% $1.69 38,739
Claude Sonnet 5 29.4% $1.79 113,409
MiMo-V2.6 Flash 28.5% $0.05 63,189
DeepSeek V4 Flash 0731 28.1% $0.16 76,525
Step 5 Preview 27.3% $0.55 84,980
DeepSeek V4 Pro 0813 26.8% $0.73 109,118
DeepSeek V4.1 Flash 26.6% $0.13 80,656
Muse Spark Open 26.3% $0.88 18,601
Claude Sonnet 4.6 25.8% $1.19 38,465
DeepSeek-V4-Pro 25.2% $0.72 110,422
Kimi K3.1 Pro 24.9% $3.64 180,052
Grok 4.5 24.9% $0.47 16,998
GPT-6.2 Sol 24.8% $2.40 120,431
GPT-5.6 Luna 24.8% $0.47 271,107
Claude Opus 4.7 23.7% $1.27 21,092
Gemini 3.7 Flash 22.8% $0.48 20,337
GPT-6.2 Astra 21.8% $7.90 109,872
GPT-5.5 21.8% $4.07 66,972
GPT-6.1 Sol 21.8% $1.32 65,893
Gemini 3.5 Flash 21.6% $0.86 30,433
GPT-6 Astra 20.7% $5.80 62,346
Ling-3.0-flash-Fin 20.6% $0.03 43,263
MiniMax-M3 20.6% $0.16 24,976
GPT-6 Luna 19.9% $0.19 214,022
Mercury 3 18.2% $0.05 73,519
Gemini 3.6 Flash 18% $0.36 16,362
GPT-6.1 Terra 17.1% $2.16 239,078
Qwen3.7-Max 15.8% $0.44 18,741
Inkling-Small 15.4% $0.09 38,499
Inkling 15.3% $0.35 37,668
GPT-6 Sol 15% $2.28 97,209
Kimi K2.6 12.4% $0.57 45,053
Gemini 3 Flash 11.8% $0.10 16,306
Nemotron 3.5 Super VL 11.6% $0.05 30,297
MiMo-V2.5-Pro 11.5% $0.04 12,625
Qwen3.7-Plus 10.5% $0.09 22,236
Gemini 3.1 Pro 9.29% $0.30 9,861
GPT-5.4 Mini 8.81% $0.81 137,206
Beam 1.0 8.81% 20,848
MiMo-V2.5 7.77% $0.02 15,557
Qwen 3.6 Plus 7.68% $0.30 17,531
Gemini 3.5 Flash-Lite 7.25% $0.08 9,682
Grok 4.20 reasoning 6.95% $0.12 8,326
MiniMax-M2.7 6.87% $0.12 14,118
Grok 4.3 6.18% $0.12 9,232
Claude Haiku 4.5 thinking 5.86% $0.23 14,371
GPT-5.4 Nano 4.97% $0.07 23,086
Mistral Medium 3.5 4.48% $1.44 36,331
Gemini 3.1 Flash-Lite 3.11% $0.01 4,699
Nemotron 3.5 Lightning 30B-A3B 2.59% $0.03 116,422
Mercury 2.5 2.49% $0.02 8,007

3 models have held number one on the official leaderboard since its release in September 2026 3 models have held number one with claims included since its release in September 2026

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

Anthropic Google DeepMind
Own leaderboard 3 leaders

Source: Official Tax Agent Bench 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 Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Gemini 4 Hydrogen Claude Opus 5 Claude Fable 5 Claude Opus 4.8 Claude Sonnet 4.6 Gemini 3 Flash Claude Haiku 4.5 0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Gemini 4 Hydrogen Claude Fable 5

Sources: Official Tax Agent Bench leaderboard and labs’ model cards

2 labs put it in a launch post in the last six months

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

Thinking Machines
Google DeepMind
DateLaunch postScoreAs the post writes it
Nov 12, 2026 Gemini 4 Hydrogen 52.3 Vals Tax Agent Bench (#1 overall)
Nov 5, 2026 Inkling 2 Small 65.8 Tax Agent Bench, 65.8%

One system card reports it

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
Nov 20, 2026 Inkling 2 Mini model card 14.1 Tax Agent Bench Capabilities

2 lab claims next to the leaderboard

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. 2 of them match, which usually means one side copied the other. Every benchmark.

Model Lab says Board says Gap
Inkling 2 Small Weighted pass 65.8% Tax Agent Bench, 65.8% 65.8% Leaderboard Same
Gemini 4 Hydrogen 52.3% Vals Tax Agent Bench (#1 overall) 52.3% Leaderboard Same

More in Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

All 12 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (88 entries), launch posts (2 scores) and system cards (1 score).
  • Pass Models scored 88 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Inkling 2 Mini.
  • Pass Frontier models 10 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Pass Headroom The best is 52.3%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on vals.ai: 88 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 18th, 2026.
  • Pass Read worked The last read worked.
Consistency 2 pass
  • Pass Lab claims Both pairs agree within 10 points.
  • Pass Known issues No published audit found problems.

How we check