Benchmark Almanac

BFCL V4

Qwen3.7-Max 75% leads since May 2026
Claude Opus 4.5 77.5% leads since November 2025
Claude Opus 4.5 77.5% leads since November 2025
In use

Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes.

V4 added agentic web search and memory categories and changed the overall formula, so its scores do not compare with V3.

Released
July 2025
Kind
Function calling
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Strongest model tested
Qwen3.7-Max #49 on the capability index
Last in a launch post
October 2026
Last model added
October 2026
Scores refreshed
November 12th, 2026
30
Models scored in lab cards
31
Launch posts list it, from 12 labs
11
System cards report it, from 7 labs

Qwen3.7-Max leads with 75%

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

Claude Opus 4.5 leads with 77.5% across all independent runs

Best score per model across the official leaderboard, Epoch AI’s own runs and any board Epoch AI copied. Each bar and row names its source.

Claude Opus 4.5 leads with 77.5%, 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.

30 models. Scroll sideways for the rest.

  1. 75.0 Qwen3.7-Max
  2. 73.4 Seed2.0 Pro
  3. 72.9 Qwen3.5 397B-A17B
  4. 72.9 Qwen3.7-Plus
  5. 72.2 Qwen3.5-122B-A10B
  6. 71.8 Spark-X2.6-4B
  7. 71.4 LFM3-1.7B
  8. 70.6 Ling-2.6-1T
  9. 70.1 Trinity Large Thinking
  10. 68.5 Qwen3.5-27B
  11. 67.7 Qwen3-Max
  12. 67.3 Qwen3.5-35B-A3B
  13. 66.5 ERNIE 5.0
  14. 66.1 Qwen3.5-9B
  15. 65.7 Keye-VL-2.0-30B-A3B
  16. 65.0 LFM2.5-24B-A2B
  17. 63.3 Qwen3.5-Omni-Plus
  18. 61.6 Nova 2.0 Pro Preview
  19. 61.4 Granite 4.2 30B
  20. 60.3 Nova 2 Lite
  21. 58.3 Nova 2 Omni Preview
  22. 57.2 Seed1.8
  23. 56.9 LFM2.5-2.6B
  24. 53.8 Nemotron 3 Nano 30B-A3B
  25. 52.9 Nemotron-Cascade 2 30B-A3B
  26. 50.3 Qwen3.5-4B
  27. 48.5 LFM2.5-8B-A1B
  28. 43.6 Qwen3.5-2B
  29. 32.5 LFM2.5-VL-3B
  30. 25.3 Qwen3.5-0.8B
  1. 1 Qwen3.7-Max 75%
  2. 2 Seed2.0 Pro 73.4%
  3. 3 Qwen3.5 397B-A17B 72.9%
  4. 4 Qwen3.7-Plus 72.9%
  5. 5 Qwen3.5-122B-A10B 72.2%
  6. 6 Spark-X2.6-4B 71.8%
  7. 7 LFM3-1.7B 71.4%
  8. 8 Ling-2.6-1T 70.6%
  9. 9 Trinity Large Thinking 70.1%
  10. 10 Qwen3.5-27B 68.5%
  11. 11 Qwen3-Max 67.7%
  12. 12 Qwen3.5-35B-A3B 67.3%
  13. 13 ERNIE 5.0 66.5%
  14. 14 Qwen3.5-9B 66.1%
  15. 15 Keye-VL-2.0-30B-A3B 65.7%
  16. 16 LFM2.5-24B-A2B 65%
  17. 17 Qwen3.5-Omni-Plus 63.3%
  18. 18 Nova 2.0 Pro Preview 61.6%
  19. 19 Granite 4.2 30B 61.4%
  20. 20 Nova 2 Lite 60.3%
  21. 21 Nova 2 Omni Preview 58.3%
  22. 22 Seed1.8 57.2%
  23. 23 LFM2.5-2.6B 56.9%
  24. 24 Nemotron 3 Nano 30B-A3B 53.8%
  25. 25 Nemotron-Cascade 2 30B-A3B 52.9%
  26. 26 Qwen3.5-4B 50.3%
  27. 27 LFM2.5-8B-A1B 48.5%
  28. 28 Qwen3.5-2B 43.6%
  29. 29 LFM2.5-VL-3B 32.5%
  30. 30 Qwen3.5-0.8B 25.3%

Source: Labs’ model cards

65 models. Scroll sideways for the rest.

  1. 77.5 Claude Opus 4.5
  2. 73.2 Claude Sonnet 4.5
  3. 72.5 Gemini 3 Pro
  4. 72.4 GLM-4.6
  5. 69.6 Grok 4.1 Fast
  6. 68.7 Claude Haiku 4.5
  7. 63.0 o3
  8. 63.0 Grok 4
  9. 59.1 Kimi K2
  10. 57.1 Command A Reasoning
  11. 56.7 DeepSeek-V3.2-Exp
  12. 56.2 Gemini 2.5 Flash
  13. 55.9 GPT-5.2
  14. 55.5 GPT-5 mini
  15. 54.7 xLAM-2-32b-fc-r
  16. 54.0 GPT-4.1
  17. 53.2 o4-mini
  18. 53.1 Llama-xLAM-2-70b-fc-r
  19. 52.1 Qwen3-235B-A22B-Instruct
  20. 51.5 GPT-5 nano
  21. 50.5 GPT-4.1 mini
  22. 48.7 Qwen3-32B
  23. 46.7 Llama-xLAM-2-8b-fc-r
  24. 46.5 Cohere Command A
  25. 42.6 Qwen3-8B
  26. 41.4 Qwen3-30B-A3B-Instruct
  27. 41.2 xLAM-2-3b-fc-r
  28. 41.0 Qwen3-14B
  29. 38.4 Mistral Large 2
  30. 37.7 Mistral Medium 3
  31. 37.3 Llama 4 Maverick
  32. 37.1 Mistral Small 3.2
  33. 36.9 Gemini 2.5 Flash-Lite
  34. 35.7 qwen3-4b-instruct-2507
  35. 33.0 GPT-4.1 nano
  36. 32.1 Command R7B
  37. 31.9 Llama 3.3 70B Instruct
  38. 30.4 xLAM-2-1b-fc-r
  39. 30.4 Gemma 3 12B
  40. 29.5 Gemma 3 27B
  41. 28.8 Phi-4
  42. 28.4 Qwen3-1.7B
  43. 28.1 Llama 4 Scout
  44. 27.9 Palmyra X4
  45. 27.6 Mistral NeMo
  46. 27.1 Granite-3.1-8B-Instruct
  47. 27.1 Nova 2 Lite
  48. 27.0 Falcon3-10B-Instruct
  49. 26.9 Granite 3.2 8B Instruct
  50. 25.8 Llama 3.1-8B
  51. 25.6 MiniCPM3-4B
  52. 25.0 Amazon Nova Pro
  53. 24.0 Falcon3-7B-Instruct
  54. 23.9 Qwen3-0.6B
  55. 23.2 Granite-20B-FunctionCalling
  56. 22.3 Amazon Nova Micro
  57. 21.9 Llama 3.2 3B
  58. 19.6 Gemma 3 4B
  59. 19.0 granite-4.0-350m
  60. 16.3 Falcon3-3B-Instruct
  61. 11.1 Ministral 8B
  62. 11.1 Falcon3-1B-Instruct
  63. 10.8 Llama 3.2 1B
  64. 10.0 Llama-3.1-Nemotron-Ultra-253B-v1
  65. 7.17 Gemma 3 1B
  1. 1 Claude Opus 4.5 77.5%
  2. 2 Claude Sonnet 4.5 73.2%
  3. 3 Gemini 3 Pro 72.5%
  4. 4 GLM-4.6 thinking 72.4%
  5. 5 Grok 4.1 Fast reasoning 69.6%
  6. 6 Claude Haiku 4.5 68.7%
  7. 7 o3 63%
  8. 8 Grok 4 63%
  9. 9 Kimi K2 Jul 2025 59.1%
  10. 10 Command A Reasoning 57.1%
  11. 11 DeepSeek-V3.2-Exp thinking 56.7%
  12. 12 Gemini 2.5 Flash Sep 2025 56.2%
  13. 13 GPT-5.2 55.9%
  14. 14 GPT-5 mini 55.5%
  15. 15 xLAM-2-32b-fc-r 54.7%
  16. 16 GPT-4.1 54%
  17. 17 o4-mini 53.2%
  18. 18 Llama-xLAM-2-70b-fc-r 53.1%
  19. 19 Qwen3-235B-A22B-Instruct Jul 2025 52.1%
  20. 20 GPT-5 nano 51.5%
  21. 21 GPT-4.1 mini 50.5%
  22. 22 Qwen3-32B 48.7%
  23. 23 Llama-xLAM-2-8b-fc-r 46.7%
  24. 24 Cohere Command A 46.5%
  25. 25 Qwen3-8B 42.6%
  26. 26 Qwen3-30B-A3B-Instruct Jul 2025 41.4%
  27. 27 xLAM-2-3b-fc-r 41.2%
  28. 28 Qwen3-14B 41%
  29. 29 Mistral Large 2 Nov 2024 38.4%
  30. 30 Mistral Medium 3 37.7%
  31. 31 Llama 4 Maverick 37.3%
  32. 32 Mistral Small 3.2 37.1%
  33. 33 Gemini 2.5 Flash-Lite Sep 2025 36.9%
  34. 34 qwen3-4b-instruct-2507 35.7%
  35. 35 GPT-4.1 nano 33%
  36. 36 Command R7B 32.1%
  37. 37 Llama 3.3 70B Instruct 31.9%
  38. 38 xLAM-2-1b-fc-r 30.4%
  39. 39 Gemma 3 12B 30.4%
  40. 40 Gemma 3 27B 29.5%
  41. 41 Phi-4 28.8%
  42. 42 Qwen3-1.7B 28.4%
  43. 43 Llama 4 Scout 28.1%
  44. 44 Palmyra X4 27.9%
  45. 45 Mistral NeMo 27.6%
  46. 46 Granite-3.1-8B-Instruct 27.1%
  47. 47 Nova 2 Lite 27.1%
  48. 48 Falcon3-10B-Instruct 27%
  49. 49 Granite 3.2 8B Instruct 26.9%
  50. 50 Llama 3.1-8B 25.8%
  51. 51 MiniCPM3-4B 25.6%
  52. 52 Amazon Nova Pro 25%
  53. 53 Falcon3-7B-Instruct 24%
  54. 54 Qwen3-0.6B 23.9%
  55. 55 Granite-20B-FunctionCalling 23.2%
  56. 56 Amazon Nova Micro 22.3%
  57. 57 Llama 3.2 3B 21.9%
  58. 58 Gemma 3 4B 19.6%
  59. 59 granite-4.0-350m 19%
  60. 60 Falcon3-3B-Instruct 16.3%
  61. 61 Ministral 8B 11.1%
  62. 62 Falcon3-1B-Instruct 11.1%
  63. 63 Llama 3.2 1B 10.8%
  64. 64 Llama-3.1-Nemotron-Ultra-253B-v1 10%
  65. 65 Gemma 3 1B 7.17%

Source: Official BFCL V4 leaderboard

94 models. Scroll sideways for the rest.

  1. 77.5 Claude Opus 4.5
  2. 75.0 Qwen3.7-Max
  3. 73.4 Seed2.0 Pro
  4. 73.2 Claude Sonnet 4.5
  5. 72.9 Qwen3.5 397B-A17B
  6. 72.9 Qwen3.7-Plus
  7. 72.5 Gemini 3 Pro
  8. 72.4 GLM-4.6
  9. 72.2 Qwen3.5-122B-A10B
  10. 71.8 Spark-X2.6-4B
  11. 71.4 LFM3-1.7B
  12. 70.6 Ling-2.6-1T
  13. 70.1 Trinity Large Thinking
  14. 69.6 Grok 4.1 Fast
  15. 68.7 Claude Haiku 4.5
  16. 68.5 Qwen3.5-27B
  17. 67.7 Qwen3-Max
  18. 67.3 Qwen3.5-35B-A3B
  19. 66.5 ERNIE 5.0
  20. 66.1 Qwen3.5-9B
  21. 65.7 Keye-VL-2.0-30B-A3B
  22. 65.0 LFM2.5-24B-A2B
  23. 63.3 Qwen3.5-Omni-Plus
  24. 63.0 o3
  25. 63.0 Grok 4
  26. 61.6 Nova 2.0 Pro Preview
  27. 61.4 Granite 4.2 30B
  28. 60.3 Nova 2 Lite
  29. 59.1 Kimi K2
  30. 58.3 Nova 2 Omni Preview
  31. 57.2 Seed1.8
  32. 57.1 Command A Reasoning
  33. 56.9 LFM2.5-2.6B
  34. 56.7 DeepSeek-V3.2-Exp
  35. 56.2 Gemini 2.5 Flash
  36. 55.9 GPT-5.2
  37. 55.5 GPT-5 mini
  38. 54.7 xLAM-2-32b-fc-r
  39. 54.0 GPT-4.1
  40. 53.8 Nemotron 3 Nano 30B-A3B
  41. 53.2 o4-mini
  42. 53.1 Llama-xLAM-2-70b-fc-r
  43. 52.9 Nemotron-Cascade 2 30B-A3B
  44. 52.1 Qwen3-235B-A22B-Instruct
  45. 51.5 GPT-5 nano
  46. 50.5 GPT-4.1 mini
  47. 50.3 Qwen3.5-4B
  48. 48.7 Qwen3-32B
  49. 48.5 LFM2.5-8B-A1B
  50. 46.7 Llama-xLAM-2-8b-fc-r
  51. 46.5 Cohere Command A
  52. 43.6 Qwen3.5-2B
  53. 42.6 Qwen3-8B
  54. 41.4 Qwen3-30B-A3B-Instruct
  55. 41.2 xLAM-2-3b-fc-r
  56. 41.0 Qwen3-14B
  57. 38.4 Mistral Large 2
  58. 37.7 Mistral Medium 3
  59. 37.3 Llama 4 Maverick
  60. 37.1 Mistral Small 3.2
  61. 36.9 Gemini 2.5 Flash-Lite
  62. 35.7 qwen3-4b-instruct-2507
  63. 33.0 GPT-4.1 nano
  64. 32.5 LFM2.5-VL-3B
  65. 32.1 Command R7B
  66. 31.9 Llama 3.3 70B Instruct
  67. 30.4 xLAM-2-1b-fc-r
  68. 30.4 Gemma 3 12B
  69. 29.5 Gemma 3 27B
  70. 28.8 Phi-4
  71. 28.4 Qwen3-1.7B
  72. 28.1 Llama 4 Scout
  73. 27.9 Palmyra X4
  74. 27.6 Mistral NeMo
  75. 27.1 Granite-3.1-8B-Instruct
  76. 27.0 Falcon3-10B-Instruct
  77. 26.9 Granite 3.2 8B Instruct
  78. 25.8 Llama 3.1-8B
  79. 25.6 MiniCPM3-4B
  80. 25.3 Qwen3.5-0.8B
  81. 25.0 Amazon Nova Pro
  82. 24.0 Falcon3-7B-Instruct
  83. 23.9 Qwen3-0.6B
  84. 23.2 Granite-20B-FunctionCalling
  85. 22.3 Amazon Nova Micro
  86. 21.9 Llama 3.2 3B
  87. 19.6 Gemma 3 4B
  88. 19.0 granite-4.0-350m
  89. 16.3 Falcon3-3B-Instruct
  90. 11.1 Ministral 8B
  91. 11.1 Falcon3-1B-Instruct
  92. 10.8 Llama 3.2 1B
  93. 10.0 Llama-3.1-Nemotron-Ultra-253B-v1
  94. 7.17 Gemma 3 1B
  1. 1 Claude Opus 4.5 77.5%
  2. 2 Qwen3.7-Max 75%
  3. 3 Seed2.0 Pro 73.4%
  4. 4 Claude Sonnet 4.5 73.2%
  5. 5 Qwen3.5 397B-A17B 72.9%
  6. 6 Qwen3.7-Plus 72.9%
  7. 7 Gemini 3 Pro 72.5%
  8. 8 GLM-4.6 thinking 72.4%
  9. 9 Qwen3.5-122B-A10B 72.2%
  10. 10 Spark-X2.6-4B 71.8%
  11. 11 LFM3-1.7B 71.4%
  12. 12 Ling-2.6-1T 70.6%
  13. 13 Trinity Large Thinking 70.1%
  14. 14 Grok 4.1 Fast reasoning 69.6%
  15. 15 Claude Haiku 4.5 68.7%
  16. 16 Qwen3.5-27B 68.5%
  17. 17 Qwen3-Max 67.7%
  18. 18 Qwen3.5-35B-A3B 67.3%
  19. 19 ERNIE 5.0 66.5%
  20. 20 Qwen3.5-9B 66.1%
  21. 21 Keye-VL-2.0-30B-A3B 65.7%
  22. 22 LFM2.5-24B-A2B 65%
  23. 23 Qwen3.5-Omni-Plus 63.3%
  24. 24 o3 63%
  25. 25 Grok 4 63%
  26. 26 Nova 2.0 Pro Preview 61.6%
  27. 27 Granite 4.2 30B 61.4%
  28. 28 Nova 2 Lite 60.3%
  29. 29 Kimi K2 Jul 2025 59.1%
  30. 30 Nova 2 Omni Preview 58.3%
  31. 31 Seed1.8 57.2%
  32. 32 Command A Reasoning 57.1%
  33. 33 LFM2.5-2.6B 56.9%
  34. 34 DeepSeek-V3.2-Exp thinking 56.7%
  35. 35 Gemini 2.5 Flash Sep 2025 56.2%
  36. 36 GPT-5.2 55.9%
  37. 37 GPT-5 mini 55.5%
  38. 38 xLAM-2-32b-fc-r 54.7%
  39. 39 GPT-4.1 54%
  40. 40 Nemotron 3 Nano 30B-A3B 53.8%
  41. 41 o4-mini 53.2%
  42. 42 Llama-xLAM-2-70b-fc-r 53.1%
  43. 43 Nemotron-Cascade 2 30B-A3B 52.9%
  44. 44 Qwen3-235B-A22B-Instruct Jul 2025 52.1%
  45. 45 GPT-5 nano 51.5%
  46. 46 GPT-4.1 mini 50.5%
  47. 47 Qwen3.5-4B 50.3%
  48. 48 Qwen3-32B 48.7%
  49. 49 LFM2.5-8B-A1B 48.5%
  50. 50 Llama-xLAM-2-8b-fc-r 46.7%
  51. 51 Cohere Command A 46.5%
  52. 52 Qwen3.5-2B 43.6%
  53. 53 Qwen3-8B 42.6%
  54. 54 Qwen3-30B-A3B-Instruct Jul 2025 41.4%
  55. 55 xLAM-2-3b-fc-r 41.2%
  56. 56 Qwen3-14B 41%
  57. 57 Mistral Large 2 Nov 2024 38.4%
  58. 58 Mistral Medium 3 37.7%
  59. 59 Llama 4 Maverick 37.3%
  60. 60 Mistral Small 3.2 37.1%
  61. 61 Gemini 2.5 Flash-Lite Sep 2025 36.9%
  62. 62 qwen3-4b-instruct-2507 35.7%
  63. 63 GPT-4.1 nano 33%
  64. 64 LFM2.5-VL-3B 32.5%
  65. 65 Command R7B 32.1%
  66. 66 Llama 3.3 70B Instruct 31.9%
  67. 67 xLAM-2-1b-fc-r 30.4%
  68. 68 Gemma 3 12B 30.4%
  69. 69 Gemma 3 27B 29.5%
  70. 70 Phi-4 28.8%
  71. 71 Qwen3-1.7B 28.4%
  72. 72 Llama 4 Scout 28.1%
  73. 73 Palmyra X4 27.9%
  74. 74 Mistral NeMo 27.6%
  75. 75 Granite-3.1-8B-Instruct 27.1%
  76. 76 Falcon3-10B-Instruct 27%
  77. 77 Granite 3.2 8B Instruct 26.9%
  78. 78 Llama 3.1-8B 25.8%
  79. 79 MiniCPM3-4B 25.6%
  80. 80 Qwen3.5-0.8B 25.3%
  81. 81 Amazon Nova Pro 25%
  82. 82 Falcon3-7B-Instruct 24%
  83. 83 Qwen3-0.6B 23.9%
  84. 84 Granite-20B-FunctionCalling 23.2%
  85. 85 Amazon Nova Micro 22.3%
  86. 86 Llama 3.2 3B 21.9%
  87. 87 Gemma 3 4B 19.6%
  88. 88 granite-4.0-350m 19%
  89. 89 Falcon3-3B-Instruct 16.3%
  90. 90 Ministral 8B 11.1%
  91. 91 Falcon3-1B-Instruct 11.1%
  92. 92 Llama 3.2 1B 10.8%
  93. 93 Llama-3.1-Nemotron-Ultra-253B-v1 10%
  94. 94 Gemma 3 1B 7.17%

Sources: Official BFCL V4 leaderboard and labs’ model cards

#ModelScoreSettingDateSource
1 Qwen3.7-Max 75% May 2026 Lab model card
2 Seed2.0 Pro 73.4% Feb 2026 Lab model card
3 Qwen3.5 397B-A17B 72.9% Feb 2026 Lab model card
4 Qwen3.7-Plus 72.9% Jun 2026 Lab model card
5 Qwen3.5-122B-A10B 72.2% Feb 2026 Lab model card
6 Spark-X2.6-4B 71.8% Oct 2026 Lab model card
7 LFM3-1.7B 71.4% Oct 2026 Lab model card
8 Ling-2.6-1T 70.6% Apr 2026 Lab model card
9 Trinity Large Thinking 70.1% Apr 2026 Lab model card
10 Qwen3.5-27B 68.5% Feb 2026 Lab model card
11 Qwen3-Max 67.7% Jan 2026 Lab model card
12 Qwen3.5-35B-A3B 67.3% Feb 2026 Lab model card
13 ERNIE 5.0 66.5% Feb 2026 Lab model card
14 Qwen3.5-9B 66.1% Mar 2026 Lab model card
15 Keye-VL-2.0-30B-A3B 65.7% May 2026 Lab model card
16 LFM2.5-24B-A2B 65% Oct 2026 Lab model card
17 Qwen3.5-Omni-Plus 63.3% Mar 2026 Lab model card
18 Nova 2.0 Pro Preview 61.6% Dec 2025 Lab model card
19 Granite 4.2 30B 61.4% Aug 2026 Lab model card
20 Nova 2 Lite 60.3% Dec 2025 Lab model card
21 Nova 2 Omni Preview 58.3% Dec 2025 Lab model card
22 Seed1.8 57.2% Dec 2025 Lab model card
23 LFM2.5-2.6B 56.9% Aug 2026 Lab model card
24 Nemotron 3 Nano 30B-A3B 53.8% Dec 2025 Lab model card
25 Nemotron-Cascade 2 30B-A3B 52.9% Mar 2026 Lab model card
26 Qwen3.5-4B 50.3% Mar 2026 Lab model card
27 LFM2.5-8B-A1B 48.5% May 2026 Lab model card
28 Qwen3.5-2B 43.6% Mar 2026 Lab model card
29 LFM2.5-VL-3B 32.5% Aug 2026 Lab model card
30 Qwen3.5-0.8B 25.3% Mar 2026 Lab model card
#ModelScoreSettingDateSource
1 Claude Opus 4.5 77.5% native function calling Nov 2025 Official leaderboard
2 Claude Sonnet 4.5 73.2% native function calling Sep 2025 Official leaderboard
3 Gemini 3 Pro 72.5% prompted Nov 2025 Official leaderboard
4 GLM-4.6 72.4% thinking, native function calling Sep 2025 Official leaderboard
5 Grok 4.1 Fast 69.6% reasoning, native function calling Nov 2025 Official leaderboard
6 Claude Haiku 4.5 68.7% native function calling Oct 2025 Official leaderboard
7 o3 63% prompted Apr 2025 Official leaderboard
8 Grok 4 63% prompted Jul 2025 Official leaderboard
9 Kimi K2 Jul 2025 59.1% native function calling Jul 2025 Official leaderboard
10 Command A Reasoning 57.1% native function calling Aug 2025 Official leaderboard
11 DeepSeek-V3.2-Exp 56.7% thinking, prompted Sep 2025 Official leaderboard
12 Gemini 2.5 Flash Sep 2025 56.2% native function calling Sep 2025 Official leaderboard
13 GPT-5.2 55.9% native function calling Dec 2025 Official leaderboard
14 GPT-5 mini 55.5% native function calling Aug 2025 Official leaderboard
15 xLAM-2-32b-fc-r 54.7% native function calling Mar 2025 Official leaderboard
16 GPT-4.1 54% native function calling Apr 2025 Official leaderboard
17 o4-mini 53.2% native function calling Apr 2025 Official leaderboard
18 Llama-xLAM-2-70b-fc-r 53.1% native function calling Mar 2025 Official leaderboard
19 Qwen3-235B-A22B-Instruct Jul 2025 52.1% prompted Jul 2025 Official leaderboard
20 GPT-5 nano 51.5% native function calling Aug 2025 Official leaderboard
21 GPT-4.1 mini 50.5% native function calling Apr 2025 Official leaderboard
22 Qwen3-32B 48.7% native function calling Apr 2025 Official leaderboard
23 Llama-xLAM-2-8b-fc-r 46.7% native function calling Mar 2025 Official leaderboard
24 Cohere Command A 46.5% native function calling Mar 2025 Official leaderboard
25 Qwen3-8B 42.6% native function calling Apr 2025 Official leaderboard
26 Qwen3-30B-A3B-Instruct Jul 2025 41.4% native function calling Jul 2025 Official leaderboard
27 xLAM-2-3b-fc-r 41.2% native function calling Mar 2025 Official leaderboard
28 Qwen3-14B 41% native function calling Apr 2025 Official leaderboard
29 Mistral Large 2 Nov 2024 38.4% native function calling Nov 2024 Official leaderboard
30 Mistral Medium 3 37.7% May 2025 Official leaderboard
31 Llama 4 Maverick 37.3% native function calling Apr 2025 Official leaderboard
32 Mistral Small 3.2 37.1% native function calling Jun 2025 Official leaderboard
33 Gemini 2.5 Flash-Lite Sep 2025 36.9% native function calling Sep 2025 Official leaderboard
34 qwen3-4b-instruct-2507 35.7% native function calling Aug 2025 Official leaderboard
35 GPT-4.1 nano 33% native function calling Apr 2025 Official leaderboard
36 Command R7B 32.1% native function calling Dec 2024 Official leaderboard
37 Llama 3.3 70B Instruct 31.9% native function calling Dec 2024 Official leaderboard
38 xLAM-2-1b-fc-r 30.4% native function calling Mar 2025 Official leaderboard
39 Gemma 3 12B 30.4% prompted Mar 2025 Official leaderboard
40 Gemma 3 27B 29.5% prompted Mar 2025 Official leaderboard
41 Phi-4 28.8% prompted Dec 2024 Official leaderboard
42 Qwen3-1.7B 28.4% native function calling Apr 2025 Official leaderboard
43 Llama 4 Scout 28.1% native function calling Apr 2025 Official leaderboard
44 Palmyra X4 27.9% native function calling Oct 2024 Official leaderboard
45 Mistral NeMo 27.6% native function calling Jul 2024 Official leaderboard
46 Granite-3.1-8B-Instruct 27.1% native function calling Dec 2024 Official leaderboard
47 Nova 2 Lite 27.1% native function calling Dec 2025 Official leaderboard
48 Falcon3-10B-Instruct 27% native function calling Dec 2024 Official leaderboard
49 Granite 3.2 8B Instruct 26.9% native function calling Feb 2025 Official leaderboard
50 Llama 3.1-8B 25.8% prompted Jul 2024 Official leaderboard
51 MiniCPM3-4B 25.6% native function calling Sep 2024 Official leaderboard
52 Amazon Nova Pro 25% native function calling Dec 2024 Official leaderboard
53 Falcon3-7B-Instruct 24% native function calling Dec 2024 Official leaderboard
54 Qwen3-0.6B 23.9% native function calling Apr 2025 Official leaderboard
55 Granite-20B-FunctionCalling 23.2% native function calling Jul 2024 Official leaderboard
56 Amazon Nova Micro 22.3% native function calling Dec 2024 Official leaderboard
57 Llama 3.2 3B 21.9% native function calling Sep 2024 Official leaderboard
58 Gemma 3 4B 19.6% prompted Mar 2025 Official leaderboard
59 granite-4.0-350m 19% native function calling Oct 2025 Official leaderboard
60 Falcon3-3B-Instruct 16.3% native function calling Dec 2024 Official leaderboard
61 Ministral 8B 11.1% native function calling Oct 2024 Official leaderboard
62 Falcon3-1B-Instruct 11.1% native function calling Dec 2024 Official leaderboard
63 Llama 3.2 1B 10.8% native function calling Sep 2024 Official leaderboard
64 Llama-3.1-Nemotron-Ultra-253B-v1 10% native function calling Apr 2025 Official leaderboard
65 Gemma 3 1B 7.17% prompted Mar 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 4.5 77.5% native function calling Nov 2025 Official leaderboard
2 Qwen3.7-Max 75% May 2026 Lab model card
3 Seed2.0 Pro 73.4% Feb 2026 Lab model card
4 Claude Sonnet 4.5 73.2% native function calling Sep 2025 Official leaderboard
5 Qwen3.5 397B-A17B 72.9% Feb 2026 Lab model card
6 Qwen3.7-Plus 72.9% Jun 2026 Lab model card
7 Gemini 3 Pro 72.5% prompted Nov 2025 Official leaderboard
8 GLM-4.6 72.4% thinking, native function calling Sep 2025 Official leaderboard
9 Qwen3.5-122B-A10B 72.2% Feb 2026 Lab model card
10 Spark-X2.6-4B 71.8% Oct 2026 Lab model card
11 LFM3-1.7B 71.4% Oct 2026 Lab model card
12 Ling-2.6-1T 70.6% Apr 2026 Lab model card
13 Trinity Large Thinking 70.1% Apr 2026 Lab model card
14 Grok 4.1 Fast 69.6% reasoning, native function calling Nov 2025 Official leaderboard
15 Claude Haiku 4.5 68.7% native function calling Oct 2025 Official leaderboard
16 Qwen3.5-27B 68.5% Feb 2026 Lab model card
17 Qwen3-Max 67.7% Jan 2026 Lab model card
18 Qwen3.5-35B-A3B 67.3% Feb 2026 Lab model card
19 ERNIE 5.0 66.5% Feb 2026 Lab model card
20 Qwen3.5-9B 66.1% Mar 2026 Lab model card
21 Keye-VL-2.0-30B-A3B 65.7% May 2026 Lab model card
22 LFM2.5-24B-A2B 65% Oct 2026 Lab model card
23 Qwen3.5-Omni-Plus 63.3% Mar 2026 Lab model card
24 o3 63% prompted Apr 2025 Official leaderboard
25 Grok 4 63% prompted Jul 2025 Official leaderboard
26 Nova 2.0 Pro Preview 61.6% Dec 2025 Lab model card
27 Granite 4.2 30B 61.4% Aug 2026 Lab model card
28 Nova 2 Lite 60.3% Dec 2025 Lab model card
29 Kimi K2 Jul 2025 59.1% native function calling Jul 2025 Official leaderboard
30 Nova 2 Omni Preview 58.3% Dec 2025 Lab model card
31 Seed1.8 57.2% Dec 2025 Lab model card
32 Command A Reasoning 57.1% native function calling Aug 2025 Official leaderboard
33 LFM2.5-2.6B 56.9% Aug 2026 Lab model card
34 DeepSeek-V3.2-Exp 56.7% thinking, prompted Sep 2025 Official leaderboard
35 Gemini 2.5 Flash Sep 2025 56.2% native function calling Sep 2025 Official leaderboard
36 GPT-5.2 55.9% native function calling Dec 2025 Official leaderboard
37 GPT-5 mini 55.5% native function calling Aug 2025 Official leaderboard
38 xLAM-2-32b-fc-r 54.7% native function calling Mar 2025 Official leaderboard
39 GPT-4.1 54% native function calling Apr 2025 Official leaderboard
40 Nemotron 3 Nano 30B-A3B 53.8% Dec 2025 Lab model card
41 o4-mini 53.2% native function calling Apr 2025 Official leaderboard
42 Llama-xLAM-2-70b-fc-r 53.1% native function calling Mar 2025 Official leaderboard
43 Nemotron-Cascade 2 30B-A3B 52.9% Mar 2026 Lab model card
44 Qwen3-235B-A22B-Instruct Jul 2025 52.1% prompted Jul 2025 Official leaderboard
45 GPT-5 nano 51.5% native function calling Aug 2025 Official leaderboard
46 GPT-4.1 mini 50.5% native function calling Apr 2025 Official leaderboard
47 Qwen3.5-4B 50.3% Mar 2026 Lab model card
48 Qwen3-32B 48.7% native function calling Apr 2025 Official leaderboard
49 LFM2.5-8B-A1B 48.5% May 2026 Lab model card
50 Llama-xLAM-2-8b-fc-r 46.7% native function calling Mar 2025 Official leaderboard
51 Cohere Command A 46.5% native function calling Mar 2025 Official leaderboard
52 Qwen3.5-2B 43.6% Mar 2026 Lab model card
53 Qwen3-8B 42.6% native function calling Apr 2025 Official leaderboard
54 Qwen3-30B-A3B-Instruct Jul 2025 41.4% native function calling Jul 2025 Official leaderboard
55 xLAM-2-3b-fc-r 41.2% native function calling Mar 2025 Official leaderboard
56 Qwen3-14B 41% native function calling Apr 2025 Official leaderboard
57 Mistral Large 2 Nov 2024 38.4% native function calling Nov 2024 Official leaderboard
58 Mistral Medium 3 37.7% May 2025 Official leaderboard
59 Llama 4 Maverick 37.3% native function calling Apr 2025 Official leaderboard
60 Mistral Small 3.2 37.1% native function calling Jun 2025 Official leaderboard
61 Gemini 2.5 Flash-Lite Sep 2025 36.9% native function calling Sep 2025 Official leaderboard
62 qwen3-4b-instruct-2507 35.7% native function calling Aug 2025 Official leaderboard
63 GPT-4.1 nano 33% native function calling Apr 2025 Official leaderboard
64 LFM2.5-VL-3B 32.5% Aug 2026 Lab model card
65 Command R7B 32.1% native function calling Dec 2024 Official leaderboard
66 Llama 3.3 70B Instruct 31.9% native function calling Dec 2024 Official leaderboard
67 xLAM-2-1b-fc-r 30.4% native function calling Mar 2025 Official leaderboard
68 Gemma 3 12B 30.4% prompted Mar 2025 Official leaderboard
69 Gemma 3 27B 29.5% prompted Mar 2025 Official leaderboard
70 Phi-4 28.8% prompted Dec 2024 Official leaderboard
71 Qwen3-1.7B 28.4% native function calling Apr 2025 Official leaderboard
72 Llama 4 Scout 28.1% native function calling Apr 2025 Official leaderboard
73 Palmyra X4 27.9% native function calling Oct 2024 Official leaderboard
74 Mistral NeMo 27.6% native function calling Jul 2024 Official leaderboard
75 Granite-3.1-8B-Instruct 27.1% native function calling Dec 2024 Official leaderboard
76 Falcon3-10B-Instruct 27% native function calling Dec 2024 Official leaderboard
77 Granite 3.2 8B Instruct 26.9% native function calling Feb 2025 Official leaderboard
78 Llama 3.1-8B 25.8% prompted Jul 2024 Official leaderboard
79 MiniCPM3-4B 25.6% native function calling Sep 2024 Official leaderboard
80 Qwen3.5-0.8B 25.3% Mar 2026 Lab model card
81 Amazon Nova Pro 25% native function calling Dec 2024 Official leaderboard
82 Falcon3-7B-Instruct 24% native function calling Dec 2024 Official leaderboard
83 Qwen3-0.6B 23.9% native function calling Apr 2025 Official leaderboard
84 Granite-20B-FunctionCalling 23.2% native function calling Jul 2024 Official leaderboard
85 Amazon Nova Micro 22.3% native function calling Dec 2024 Official leaderboard
86 Llama 3.2 3B 21.9% native function calling Sep 2024 Official leaderboard
87 Gemma 3 4B 19.6% prompted Mar 2025 Official leaderboard
88 granite-4.0-350m 19% native function calling Oct 2025 Official leaderboard
89 Falcon3-3B-Instruct 16.3% native function calling Dec 2024 Official leaderboard
90 Ministral 8B 11.1% native function calling Oct 2024 Official leaderboard
91 Falcon3-1B-Instruct 11.1% native function calling Dec 2024 Official leaderboard
92 Llama 3.2 1B 10.8% native function calling Sep 2024 Official leaderboard
93 Llama-3.1-Nemotron-Ultra-253B-v1 10% native function calling Apr 2025 Official leaderboard
94 Gemma 3 1B 7.17% prompted Mar 2025 Official leaderboard

The benchmark's own leaderboard has 80 entries, 65 of them models we track. Switch the scores above to rank them together.

The best score, 77.5%, costs $87 for the full run

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.

Qwen Cohere Anthropic Z.ai Cost frontier One model at each effort level
0% 20% 40% 60% 80% 100% $0.10 $1.00 $10 $100 $1,000 Cost of the full run, log scale Claude Opus 4.5 Claude Sonnet 4.5 GLM-4.6, thinking Command A Reasoning Qwen3-235B-A22B-Instruct Qwen3-14B Qwen3-0.6B 0% 20% 40% 60% 80% 100% $0.10 $1.00 $10 $100 $1,000 Cost of the full run, log scale Claude Opus 4.5 GLM-4.6, thinking

Source: Official BFCL V4 leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
Claude Opus 4.5 native function calling 77.5% $87
Claude Sonnet 4.5 native function calling 73.2% $44
Gemini 3 Pro prompted 72.5% $299
GLM-4.6 thinking native function calling 72.4% $4.64
Grok 4.1 Fast reasoning native function calling 69.6% $17
Claude Haiku 4.5 native function calling 68.7% $14
Gemini 3 Pro native function calling 68.1% $225
o3 prompted 63% $235
Grok 4 prompted 63% $348
Grok 4 native function calling 61.4% $355
Kimi K2 Jul 2025 native function calling 59.1% $6.19
Grok 4.1 Fast non-reasoning native function calling 58.3% $16
Command A Reasoning native function calling 57.1% $3.04
DeepSeek-V3.2-Exp thinking prompted 56.7% $58
Gemini 2.5 Flash Sep 2025 native function calling 56.2% $26
GPT-5.2 native function calling 55.9% $86
GPT-5 mini native function calling 55.5% $22
xLAM-2-32b-fc-r native function calling 54.7% $6.00
DeepSeek-V3.2-Exp native function calling 54.1% $6.71
GPT-4.1 native function calling 54% $101
o4-mini native function calling 53.2% $82
Llama-xLAM-2-70b-fc-r native function calling 53.1% $25
Qwen3-235B-A22B-Instruct Jul 2025 prompted 52.1% $3.12
GPT-5 nano native function calling 51.5% $8.79
Gemini 2.5 Flash Sep 2025 prompted 50.9% $33
GPT-4.1 mini native function calling 50.5% $19
o4-mini prompted 50.3% $134
Qwen3-32B native function calling 48.7% $153
o3 native function calling 48.6% $133
Qwen3-235B-A22B-Instruct Jul 2025 native function calling 48% $2.50
Qwen3-32B prompted 46.8% $200
Llama-xLAM-2-8b-fc-r native function calling 46.7% $21
Cohere Command A native function calling 46.5% $91
GPT-5.2 prompted 45.3% $165
Qwen3-8B native function calling 42.6% $43
Qwen3-30B-A3B-Instruct Jul 2025 native function calling 41.4% $5.62
xLAM-2-3b-fc-r native function calling 41.2% $3.36
Qwen3-14B native function calling 41% $3.38
Qwen3-8B prompted 40.4% $64
GPT-4.1 prompted 39.4% $146
Mistral Large 2 Nov 2024 native function calling 38.4% $116
Qwen3-14B prompted 37.8% $1.35
Mistral Medium 3 37.7% $37
Mistral Medium 3 native function calling 37.6% $19
Llama 4 Maverick native function calling 37.3% $18
Mistral Small 3.2 native function calling 37.1% $5.20
Gemini 2.5 Flash-Lite Sep 2025 native function calling 36.9% $7.55
Qwen3-30B-A3B-Instruct Jul 2025 prompted 36.7% $1.56
qwen3-4b-instruct-2507 native function calling 35.7% $6.37
qwen3-4b-instruct-2507 prompted 35.5% $54
Claude Opus 4.5 prompted 33.5% $88
GPT-4.1 nano native function calling 33% $5.66
Mistral Small 3.2 prompted 32.4% $6.91
Command R7B native function calling 32.1% $1.50
Llama 3.3 70B Instruct native function calling 31.9% $30
Mistral Large 2 Nov 2024 prompted 31.8% $232
xLAM-2-1b-fc-r native function calling 30.4% $2.79
Gemma 3 12B prompted 30.4% $11
GPT-4.1 mini prompted 29.7% $21
Gemma 3 27B prompted 29.5% $12
Phi-4 prompted 28.8% $8.72
Qwen3-1.7B native function calling 28.4% $4.33
Llama 4 Scout native function calling 28.1% $25
Gemini 2.5 Flash-Lite Sep 2025 prompted 28% $7.05
Palmyra X4 native function calling 27.9% $178
GPT-5 mini prompted 27.8% $83
Mistral NeMo native function calling 27.6% $8.12
GPT-5 nano prompted 27.6% $21
Granite-3.1-8B-Instruct native function calling 27.1% $9.32
Nova 2 Lite native function calling 27.1% $78
Falcon3-10B-Instruct native function calling 27% $53
Granite 3.2 8B Instruct native function calling 26.9% $25
Llama 3.1-8B prompted 25.8% $7.49
MiniCPM3-4B native function calling 25.6% $54
Claude Haiku 4.5 prompted 25.3% $45
Amazon Nova Pro native function calling 25% $48
Claude Sonnet 4.5 prompted 24.9% $48
GPT-4.1 nano prompted 24.9% $7.42
Falcon3-7B-Instruct native function calling 24% $74
Qwen3-0.6B native function calling 23.9% $0.46
Granite-20B-FunctionCalling native function calling 23.2% $5.23
Qwen3-0.6B prompted 22.4% $3.65
Amazon Nova Micro native function calling 22.3% $1.81
MiniCPM3-4B prompted 22.1% $30
Llama 3.2 3B native function calling 21.9% $6.20
Gemma 3 4B prompted 19.6% $4.14
Mistral NeMo prompted 19.3% $14
granite-4.0-350m native function calling 19% $1.44
Falcon3-3B-Instruct native function calling 16.3% $37
Ministral 8B native function calling 11.1% $70
Falcon3-1B-Instruct native function calling 11.1% $1.72
Llama 3.2 1B native function calling 10.8% $1.64
Llama-3.1-Nemotron-Ultra-253B-v1 native function calling 10% $0.72
Gemma 3 1B prompted 7.17% $3.40

4 models have held number one in labs’ own numbers since its release in July 2025 3 models have held number one across all independent runs since its release in July 2025 3 models have held number one with claims included since its release in July 2025

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 Amazon Qwen ByteDance Seed
Own leaderboard Not tested since Dec 2025
Claude Opus 4.5 77.5%
Lab claims 4 leaders
Qwen3.7-Max 75%

Sources: Official BFCL V4 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% 2025 2026 2027 Near the ceiling (85%) Claude Opus 4.5 Claude Sonnet 4.5 o3 xLAM-2-32b-fc-r Cohere Command A Mistral Large 2 Palmyra X4 0% 20% 40% 60% 80% 100% 2025 2026 2027 Near the ceiling (85%) Claude Opus 4.5 o3

Sources: Official BFCL V4 leaderboard and labs’ model cards

4 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.

xAI
Amazon
NVIDIA
ByteDance Seed
Qwen
Baidu
Arcee AI
Ant Group
Kuaishou
Liquid AI
IBM
iFlytek
DateLaunch postScoreAs the post writes it
Oct 27, 2026 LFM3-1.7B 71.4 BFCL V4
Oct 22, 2026 Spark-X2.6-4B 71.8 BFCL-V4
Oct 13, 2026 LFM2.5-24B-A2B 65 BFCLv4
Aug 25, 2026 Granite 4.2 30B 61.4 BFCL (v4)
Aug 12, 2026 LFM2.5-VL-3B 32.5 BFCL V4
Aug 4, 2026 LFM2.5-2.6B 56.9 BFCLv4
Jun 1, 2026 Qwen3.7-Plus 72.9 BFCL-V4
May 28, 2026 LFM2.5-8B-A1B 48.5 BFCLv4
May 25, 2026 Keye-VL-2.0-30B-A3B 65.7 BFCL-V4
May 20, 2026 Qwen3.7-Max 75 BFCL-V4
Apr 29, 2026 Ling-2.6-1T 70.6 BFCL-V4
Apr 1, 2026 Trinity Large Thinking 70.1 BCFLv4
Mar 30, 2026 Qwen3.5-Omni-Plus 63.3 BFCL-V4
Mar 19, 2026 Nemotron-Cascade 2 30B-A3B 52.9 BFCL v4
Mar 2, 2026 Qwen3.5-0.8B 25.3 BFCL-V4
Mar 2, 2026 Qwen3.5-2B 43.6 BFCL-V4
Mar 2, 2026 Qwen3.5-4B 50.3 BFCL-V4
Mar 2, 2026 Qwen3.5-9B 66.1 BFCL-V4
Feb 24, 2026 Qwen3.5-35B-A3B 67.3 BFCL-V4
Feb 24, 2026 Qwen3.5-27B 68.5 BFCL-V4
Feb 24, 2026 Qwen3.5-122B-A10B 72.2 BFCL-V4
Feb 16, 2026 Qwen3.5 397B-A17B 72.9 BFCL-V4
Feb 14, 2026 Seed2.0 Pro 73.4 BFCL-v4
Feb 6, 2026 ERNIE 5.0 66.5 BFCL v4
Jan 23, 2026 Qwen3-Max 67.7 BFCL-V4
Dec 18, 2025 Seed1.8 57.2 BFCL-v4
Dec 15, 2025 Nemotron 3 Nano 30B-A3B 53.8 BFCL v4
Dec 2, 2025 Nova 2 Omni Preview 58.3 BFCL v4 Overall FC
Dec 2, 2025 Nova 2 Lite 60.3 BFCL v4 Overall FC
Dec 2, 2025 Nova 2.0 Pro Preview 61.6 BFCL v4 Overall FC
Nov 19, 2025 Grok 4.1 Fast Berkeley Function Calling v4 Benchmark, overall accuracy (chart)

11 system cards report it, from 7 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
Jun 13, 2026 Ling-2.6-1T technical report 70.6 BFCL-v4, accuracy (Table 5) Capabilities
Jun 9, 2026 Keye-VL-2.0-30B-A3B technical report 65.7 BFCL-V4 (Table 4) Capabilities
Apr 17, 2026 Qwen3.5-Omni-Plus technical report 63.3 BFCL-V4 (Table 4) Capabilities
Mar 21, 2026 Seed1.8 model card 57.2 BFCL-v4 (Table 4) Capabilities
Mar 19, 2026 Nemotron-Cascade 2 30B-A3B technical report 52.9 BFCL v4, thinking mode, avg@1 (Table 1) Capabilities
Feb 14, 2026 Seed2.0 Pro model card 73.4 BFCL-v4 (Table 11) Capabilities
Feb 4, 2026 ERNIE 5.0 technical report 66.5 BFCL v4 (Table 2) Capabilities
Dec 23, 2025 Nemotron 3 Nano 30B-A3B technical report 53.8 BFCL v4 (Table 3) Capabilities
Dec 2, 2025 Nova 2 Omni Preview technical report 58.3 BFCL v4 Overall FC (Table 5) Capabilities
Dec 2, 2025 Nova 2 Lite technical report 60.3 BFCL v4 Overall FC (Table 2) Capabilities
Dec 2, 2025 Nova 2.0 Pro Preview technical report 61.6 BFCL v4 Overall FC (Table 2) Capabilities

One lab claim 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. Every benchmark.

Model Lab says Board says Gap
Nova 2 Lite 60.3% BFCL v4 Overall FC 27.1% Leaderboard, native function calling +33 pts

3 logged changes that move scores

Changes by the maintainers, so a score from before a change may not compare with one from after it. Newest first, each with the maintainers' own source. The list may be incomplete. What we log.

  1. Scores recomputed December 17th, 2025

    The maintainers updated leaderboard scores using checkpoint f7cf735 after backend and dataset fixes. The new scores supersede the preceding checkpoint’s results. Source

  2. Scores recomputed November 4th, 2025

    The maintainers updated leaderboard scores using checkpoint 86d0374 after dataset, backend, and grading fixes. Earlier leaderboard scores use different evaluation behavior. Source

  3. Questions corrected October 13th, 2025

    The maintainers corrected prompts and ground truths across single-turn and multi-turn categories. Evaluations now use revised questions and expected calls. Source

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

Z.ai Meta Amazon Qwen ByteDance Seed
BFCL V4 6 leaders
Qwen3.7-Max 75%
Berkeley Function Calling Leaderboard (BFCL) Passed 85% in Jul 2024, by Llama 3.1-405B
Llama 3.1-405B 88.5%
BFCL V4 4 leaders
Qwen3.7-Max 75%

Source: Labs’ model cards

Replaces Berkeley Function Calling Leaderboard (BFCL)

More in Tool use

All 144 in Tool use
  • AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state. Gemini 4 Carbon 53.9%
  • MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks. Gemini 4 Argon 90.7%
  • τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state. Qwen 3.8 Max 55.1%
  • Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
  • DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%
  • Claw-Eval In use 300 human-verified agent tasks (service orchestration, multimodal, multi-turn dialogue) audited via traces and snapshots. Rubric scores with Pass^3 consistency. Claude Sonnet 4.6 81.4%