Benchmark Almanac

LiveCodeBench v6

DeepSeek-V4-Pro 93.5% leads since April 2026
o4-mini 80.2% leads since April 2025
DeepSeek-V4-Pro 93.5% leads since April 2026
In use
Version v6

The official leaderboard hasn’t scored a model released after EXAONE 4.0 32B (July 2025). Labs report their own numbers for 80 newer models, such as DeepSeek-V4-Pro at 93.5%.

Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests.

Problems released from May 2023 to April 2025, 1,055 in all.

Made by
UC Berkeley / MIT / Cornell
Released
April 2025
Kind
Competitive programming
Level
Competition
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Code
Interaction
Single turn
Environment
Code sandbox
Items from
Exams and contests
Run by
Anyone
Based on
LiveCodeBench v5
Strongest model tested
Qwen 3.8 27B #59 on the capability index
Last in a launch post
October 2026
Last model added
October 2026
Scores refreshed
November 19th, 2026

In the authors’ words

Large Language Models (LLMs) applied to code-related applications have emerged as a prominent field, attracting significant interest from both academia and industry. However, as new and improved LLMs are developed, existing evaluation benchmarks (e.g., HumanEval, MBPP) are no longer sufficient for assessing their capabilities. In this work, we propose LiveCodeBench, a comprehensive and contamination-free evaluation of LLMs for code, which continuously collects new problems over time from contests across three competition platforms, namely LeetCode, AtCoder, and CodeForces. Notably, our benchmark also focuses on a broader range of code related capabilities, such as self-repair, code execution, and test output prediction, beyond just code generation. Currently, LiveCodeBench hosts four hundred high-quality coding problems that were published between May 2023 and May 2024. We have evaluated 18 base LLMs and 34 instruction-tuned LLMs on LiveCodeBench. We present empirical findings on contamination, holistic performance comparisons, potential overfitting in existing benchmarks as well as individual model comparisons. We will release all prompts and model completions for further community analysis, along with a general toolkit for adding new scenarios and model

Abstract of LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code (arXiv:2403.07974, CC0)
92
Models scored in lab cards
76
Launch posts list it, from 29 labs
53
System cards report it, from 23 labs

DeepSeek-V4-Pro leads with 93.5%

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

o4-mini leads with 80.2% across all independent runs

Best score per model across the official leaderboard, Epoch AI’s own runs and any board Epoch AI copied. Dotted bars are reruns by Epoch AI or another board, and each row names its source.

DeepSeek-V4-Pro leads with 93.5%, 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. Striped bars are claims, and dotted bars are reruns by Epoch AI or another board.

92 models. Scroll sideways for the rest.

  1. 93.5 DeepSeek-V4-Pro
  2. 92.6 Qwen3.8-Omni-Flash
  3. 92.4 Solar Open 2
  4. 91.9 Qwen3.8 Flash
  5. 91.6 DeepSeek-V4-Flash
  6. 90.3 Qwen 3.8 27B
  7. 89.6 Kimi K2.6
  8. 89.4 Ring-2.5-1T
  9. 89.0 Nemotron 3 Ultra
  10. 88.2 Nemotron 3.5 Super VL
  11. 87.8 Seed2.0 Pro
  12. 87.7 MAI-Thinking-1
  13. 87.6 Gemini 2.5 Deep Think
  14. 87.2 Nemotron-Cascade 2 30B-A3B
  15. 87.1 Qwen 3.6 Plus
  16. 87.0 Ring-2.6-1T
  17. 86.4 Step 3.5 Flash
  18. 85.9 Qwen3-Max-2026-01-23
  19. 85.1 MiMo-V2-Flash
  20. 85.0 Kimi K2.5
  21. 84.9 GLM-4.7
  22. 83.9 Qwen3.6 27B
  23. 83.6 Qwen3.5 397B-A17B
  24. 83.3 DeepSeek-V3.2
  25. 83.1 Kimi K2 Thinking
  26. 82.8 GLM-4.6
  27. 82.8 LongCat-Flash-Thinking-2601
  28. 81.4 EXAONE 4.5 33B
  29. 80.7 K-EXAONE
  30. 80.7 Qwen3.5-27B
  31. 80.4 Qwen 3.6 35B-A3B
  32. 80.3 Flint
  33. 80.0 Gemma 4 31B IT
  34. 79.5 Seed1.8
  35. 79.4 LongCat-Flash-Thinking
  36. 78.9 Qwen3.5-122B-A10B
  37. 78.7 Nemotron 3 Super
  38. 78.3 Ring-1T
  39. 77.7 Solar Pro 3 (March 2026 update)
  40. 77.1 Gemma 4 26B A4B
  41. 77.1 Keye-VL-2.0-30B-A3B
  42. 76.2 ERNIE 5.0
  43. 75.8 STEP3-VL-10B
  44. 75.8 Granite 4.2 30B
  45. 74.8 DeepSeek-V3.1
  46. 74.6 Qwen3.5-35B-A3B
  47. 74.3 Intern-S1-Pro
  48. 74.2 Solar Open 100B
  49. 74.1 Qwen3-235B-A22B-Thinking
  50. 74.1 DeepSeek-V3.2-Exp
  51. 73.5 LFM3-1.7B
  52. 73.3 DeepSeek-R1
  53. 72.0 Gemma 4 12B
  54. 71.7 Sarvam 105B
  55. 70.3 North Mini Code 1.0
  56. 70.1 Qwen3-VL-235B-A22B
  57. 70.0 Sarvam 30B
  58. 69.3 INTELLECT-3
  59. 69.1 DiffusionGemma 26B A4B
  60. 68.7 Qwen3-Next-80B-A3B
  61. 68.6 Falcon H1R 7B
  62. 68.3 Nemotron 3 Nano 30B-A3B
  63. 67.9 NousCoder-14B
  64. 67.4 seed-oss-36b-instruct
  65. 67.1 Step 3
  66. 66.7 EXAONE 4.0 32B
  67. 65.6 Qwen3.5-9B
  68. 65.6 Qwen3.5-Omni-Plus
  69. 65.6 Ling-2.6-1T
  70. 65.0 MiniMax-M1-80k
  71. 64.6 Ministral 3 14B
  72. 64.0 glm-4.7-flash
  73. 63.5 Code World Model (CWM)
  74. 63.4 Beam 1.0 Base
  75. 61.4 Hermes 4 405B
  76. 59.4 LFM2.5-2.6B
  77. 58.9 Qwen3-Coder-Next
  78. 56.6 MiniCPM-V-4.7-35B-A3B
  79. 55.8 Qwen3.5-4B
  80. 53.7 Kimi K2
  81. 52.6 LongCat-Flash-Omni
  82. 52.0 Gemma 4 E4B
  83. 50.3 Magistral Medium 1.0
  84. 49.3 MiMo-7B-RL
  85. 48.0 LongCat-Flash-Chat
  86. 47.4 Magistral Small 1.0
  87. 46.8 Reka Flash 3.1
  88. 44.0 Gemma 4 E2B
  89. 38.8 ERNIE-4.5-300B-A47B
  90. 33.7 Seed Diffusion Preview
  91. 30.3 Mistral Medium 3
  92. 21.0 LFM2-8B-A1B
  1. 1 DeepSeek-V4-Pro 93.5%
  2. 2 Qwen3.8-Omni-Flash 92.6%
  3. 3 Solar Open 2 92.4%
  4. 4 Qwen3.8 Flash 91.9%
  5. 5 DeepSeek-V4-Flash 91.6%
  6. 6 Qwen 3.8 27B 90.3%
  7. 7 Kimi K2.6 89.6%
  8. 8 Ring-2.5-1T 89.4%
  9. 9 Nemotron 3 Ultra 89%
  10. 10 Nemotron 3.5 Super VL 88.2%
  11. 11 Seed2.0 Pro 87.8%
  12. 12 MAI-Thinking-1 87.7%
  13. 13 Gemini 2.5 Deep Think 87.6%
  14. 14 Nemotron-Cascade 2 30B-A3B 87.2%
  15. 15 Qwen 3.6 Plus 87.1%
  16. 16 Ring-2.6-1T 87%
  17. 17 Step 3.5 Flash 86.4%
  18. 18 Qwen3-Max-2026-01-23 85.9%
  19. 19 MiMo-V2-Flash 85.1%
  20. 20 Kimi K2.5 85%
  21. 21 GLM-4.7 84.9%
  22. 22 Qwen3.6 27B 83.9%
  23. 23 Qwen3.5 397B-A17B 83.6%
  24. 24 DeepSeek-V3.2 83.3%
  25. 25 Kimi K2 Thinking 83.1%
  26. 26 GLM-4.6 82.8%
  27. 27 LongCat-Flash-Thinking-2601 82.8%
  28. 28 EXAONE 4.5 33B 81.4%
  29. 29 K-EXAONE 80.7%
  30. 30 Qwen3.5-27B 80.7%
  31. 31 Qwen 3.6 35B-A3B 80.4%
  32. 32 Flint 80.3%
  33. 33 Gemma 4 31B IT 80%
  34. 34 Seed1.8 79.5%
  35. 35 LongCat-Flash-Thinking 79.4%
  36. 36 Qwen3.5-122B-A10B 78.9%
  37. 37 Nemotron 3 Super 78.7%
  38. 38 Ring-1T 78.3%
  39. 39 Solar Pro 3 (March 2026 update) 77.7%
  40. 40 Gemma 4 26B A4B 77.1%
  41. 41 Keye-VL-2.0-30B-A3B 77.1%
  42. 42 ERNIE 5.0 76.2%
  43. 43 STEP3-VL-10B 75.8%
  44. 44 Granite 4.2 30B 75.8%
  45. 45 DeepSeek-V3.1 74.8%
  46. 46 Qwen3.5-35B-A3B 74.6%
  47. 47 Intern-S1-Pro 74.3%
  48. 48 Solar Open 100B 74.2%
  49. 49 Qwen3-235B-A22B-Thinking Jul 2025 74.1%
  50. 50 DeepSeek-V3.2-Exp 74.1%
  51. 51 LFM3-1.7B 73.5%
  52. 52 DeepSeek-R1 May 2025 73.3%
  53. 53 Gemma 4 12B 72%
  54. 54 Sarvam 105B 71.7%
  55. 55 North Mini Code 1.0 70.3%
  56. 56 Qwen3-VL-235B-A22B 70.1%
  57. 57 Sarvam 30B 70%
  58. 58 INTELLECT-3 69.3%
  59. 59 DiffusionGemma 26B A4B 69.1%
  60. 60 Qwen3-Next-80B-A3B 68.7%
  61. 61 Falcon H1R 7B 68.6%
  62. 62 Nemotron 3 Nano 30B-A3B 68.3%
  63. 63 NousCoder-14B 67.9%
  64. 64 seed-oss-36b-instruct 67.4%
  65. 65 Step 3 67.1%
  66. 66 EXAONE 4.0 32B 66.7%
  67. 67 Qwen3.5-9B 65.6%
  68. 68 Qwen3.5-Omni-Plus 65.6%
  69. 69 Ling-2.6-1T 65.6%
  70. 70 MiniMax-M1-80k 65%
  71. 71 Ministral 3 14B 64.6%
  72. 72 glm-4.7-flash 64%
  73. 73 Code World Model (CWM) 63.5%
  74. 74 Beam 1.0 Base 63.4%
  75. 75 Hermes 4 405B 61.4%
  76. 76 LFM2.5-2.6B 59.4%
  77. 77 Qwen3-Coder-Next 58.9%
  78. 78 MiniCPM-V-4.7-35B-A3B 56.6%
  79. 79 Qwen3.5-4B 55.8%
  80. 80 Kimi K2 Jul 2025 53.7%
  81. 81 LongCat-Flash-Omni 52.6%
  82. 82 Gemma 4 E4B 52%
  83. 83 Magistral Medium 1.0 50.3%
  84. 84 MiMo-7B-RL 49.3%
  85. 85 LongCat-Flash-Chat 48%
  86. 86 Magistral Small 1.0 47.4%
  87. 87 Reka Flash 3.1 46.8%
  88. 88 Gemma 4 E2B 44%
  89. 89 ERNIE-4.5-300B-A47B 38.8%
  90. 90 Seed Diffusion Preview 33.7%
  91. 91 Mistral Medium 3 30.3%
  92. 92 LFM2-8B-A1B 21%

Source: Labs’ model cards

111 models. Scroll sideways for the rest.

  1. 93.5 DeepSeek-V4-Pro
  2. 92.6 Qwen3.8-Omni-Flash
  3. 92.4 Solar Open 2
  4. 91.9 Qwen3.8 Flash
  5. 91.6 DeepSeek-V4-Flash
  6. 90.3 Qwen 3.8 27B
  7. 89.6 Kimi K2.6
  8. 89.4 Ring-2.5-1T
  9. 89.0 Nemotron 3 Ultra
  10. 88.2 Nemotron 3.5 Super VL
  11. 87.8 Seed2.0 Pro
  12. 87.7 MAI-Thinking-1
  13. 87.6 Gemini 2.5 Deep Think
  14. 87.2 Nemotron-Cascade 2 30B-A3B
  15. 87.1 Qwen 3.6 Plus
  16. 87.0 Ring-2.6-1T
  17. 86.4 Step 3.5 Flash
  18. 85.9 Qwen3-Max-2026-01-23
  19. 85.1 MiMo-V2-Flash
  20. 85.0 Kimi K2.5
  21. 84.9 GLM-4.7
  22. 83.9 Qwen3.6 27B
  23. 83.6 Qwen3.5 397B-A17B
  24. 83.3 DeepSeek-V3.2
  25. 83.1 Kimi K2 Thinking
  26. 82.8 GLM-4.6
  27. 82.8 LongCat-Flash-Thinking-2601
  28. 81.4 EXAONE 4.5 33B
  29. 80.7 K-EXAONE
  30. 80.7 Qwen3.5-27B
  31. 80.4 Qwen 3.6 35B-A3B
  32. 80.3 Flint
  33. 80.2 o4-mini
  34. 80.0 Gemma 4 31B IT
  35. 79.5 Seed1.8
  36. 79.4 LongCat-Flash-Thinking
  37. 78.9 Qwen3.5-122B-A10B
  38. 78.7 Nemotron 3 Super
  39. 78.3 Ring-1T
  40. 77.7 Solar Pro 3 (March 2026 update)
  41. 77.1 Gemma 4 26B A4B
  42. 77.1 Keye-VL-2.0-30B-A3B
  43. 76.2 ERNIE 5.0
  44. 75.8 o3
  45. 75.8 STEP3-VL-10B
  46. 75.8 Granite 4.2 30B
  47. 74.8 DeepSeek-V3.1
  48. 74.6 Qwen3.5-35B-A3B
  49. 74.3 Intern-S1-Pro
  50. 74.2 Solar Open 100B
  51. 74.1 Qwen3-235B-A22B-Thinking
  52. 74.1 DeepSeek-V3.2-Exp
  53. 73.6 Gemini 2.5 Pro
  54. 73.5 LFM3-1.7B
  55. 73.3 DeepSeek-R1
  56. 72.0 Gemma 4 12B
  57. 71.8 Gemini 2.5 Pro
  58. 71.7 Sarvam 105B
  59. 70.3 North Mini Code 1.0
  60. 70.1 Qwen3-VL-235B-A22B
  61. 70.0 EXAONE 4.0 32B
  62. 70.0 Sarvam 30B
  63. 69.8 OpenReasoning-Nemotron-32B
  64. 69.3 INTELLECT-3
  65. 69.1 DiffusionGemma 26B A4B
  66. 68.7 Qwen3-Next-80B-A3B
  67. 68.6 Falcon H1R 7B
  68. 68.3 Nemotron 3 Nano 30B-A3B
  69. 67.9 NousCoder-14B
  70. 67.4 o3-mini
  71. 67.4 seed-oss-36b-instruct
  72. 67.1 Step 3
  73. 66.8 OpenCodeReasoning-Nemotron-1.1-32B
  74. 66.7 Grok-3 mini
  75. 65.9 Qwen3-235B-A22B
  76. 65.6 Qwen3.5-9B
  77. 65.6 Qwen3.5-Omni-Plus
  78. 65.6 Ling-2.6-1T
  79. 65.0 MiniMax-M1-80k
  80. 64.6 Ministral 3 14B
  81. 64.0 glm-4.7-flash
  82. 63.5 Code World Model (CWM)
  83. 63.4 Beam 1.0 Base
  84. 61.9 Gemini 2.5 Flash
  85. 61.4 Hermes 4 405B
  86. 60.6 Gemini 2.5 Flash
  87. 59.4 LFM2.5-2.6B
  88. 58.9 Qwen3-Coder-Next
  89. 56.6 Claude Opus 4
  90. 56.6 MiniCPM-V-4.7-35B-A3B
  91. 55.9 Claude Sonnet 4
  92. 55.8 Qwen3.5-4B
  93. 53.7 Kimi K2
  94. 52.6 LongCat-Flash-Omni
  95. 52.0 Gemma 4 E4B
  96. 50.3 Magistral Medium 1.0
  97. 49.3 MiMo-7B-RL
  98. 48.0 LongCat-Flash-Chat
  99. 47.4 Magistral Small 1.0
  100. 46.8 Reka Flash 3.1
  101. 44.0 Gemma 4 E2B
  102. 38.8 ERNIE-4.5-300B-A47B
  103. 36.4 Claude 3.5 Sonnet (October 2024)
  104. 33.7 Seed Diffusion Preview
  105. 30.3 Mistral Medium 3
  106. 29.5 GPT-4o
  107. 28.7 GPT-4 Turbo
  108. 27.5 GPT-4o mini
  109. 27.2 DeepSeek-V3
  110. 21.0 LFM2-8B-A1B
  111. 20.2 Claude 3 Haiku
  1. 1 DeepSeek-V4-Pro 93.5%
  2. 2 Qwen3.8-Omni-Flash 92.6%
  3. 3 Solar Open 2 92.4%
  4. 4 Qwen3.8 Flash 91.9%
  5. 5 DeepSeek-V4-Flash 91.6%
  6. 6 Qwen 3.8 27B 90.3%
  7. 7 Kimi K2.6 89.6%
  8. 8 Ring-2.5-1T 89.4%
  9. 9 Nemotron 3 Ultra 89%
  10. 10 Nemotron 3.5 Super VL 88.2%
  11. 11 Seed2.0 Pro 87.8%
  12. 12 MAI-Thinking-1 87.7%
  13. 13 Gemini 2.5 Deep Think 87.6%
  14. 14 Nemotron-Cascade 2 30B-A3B 87.2%
  15. 15 Qwen 3.6 Plus 87.1%
  16. 16 Ring-2.6-1T 87%
  17. 17 Step 3.5 Flash 86.4%
  18. 18 Qwen3-Max-2026-01-23 85.9%
  19. 19 MiMo-V2-Flash 85.1%
  20. 20 Kimi K2.5 85%
  21. 21 GLM-4.7 84.9%
  22. 22 Qwen3.6 27B 83.9%
  23. 23 Qwen3.5 397B-A17B 83.6%
  24. 24 DeepSeek-V3.2 83.3%
  25. 25 Kimi K2 Thinking 83.1%
  26. 26 GLM-4.6 82.8%
  27. 27 LongCat-Flash-Thinking-2601 82.8%
  28. 28 EXAONE 4.5 33B 81.4%
  29. 29 K-EXAONE 80.7%
  30. 30 Qwen3.5-27B 80.7%
  31. 31 Qwen 3.6 35B-A3B 80.4%
  32. 32 Flint 80.3%
  33. 33 o4-mini high 80.2%
  34. 34 Gemma 4 31B IT 80%
  35. 35 Seed1.8 79.5%
  36. 36 LongCat-Flash-Thinking 79.4%
  37. 37 Qwen3.5-122B-A10B 78.9%
  38. 38 Nemotron 3 Super 78.7%
  39. 39 Ring-1T 78.3%
  40. 40 Solar Pro 3 (March 2026 update) 77.7%
  41. 41 Gemma 4 26B A4B 77.1%
  42. 42 Keye-VL-2.0-30B-A3B 77.1%
  43. 43 ERNIE 5.0 76.2%
  44. 44 o3 high 75.8%
  45. 45 STEP3-VL-10B 75.8%
  46. 46 Granite 4.2 30B 75.8%
  47. 47 DeepSeek-V3.1 74.8%
  48. 48 Qwen3.5-35B-A3B 74.6%
  49. 49 Intern-S1-Pro 74.3%
  50. 50 Solar Open 100B 74.2%
  51. 51 Qwen3-235B-A22B-Thinking Jul 2025 74.1%
  52. 52 DeepSeek-V3.2-Exp 74.1%
  53. 53 Gemini 2.5 Pro Jun 2025 73.6%
  54. 54 LFM3-1.7B 73.5%
  55. 55 DeepSeek-R1 May 2025 73.3%
  56. 56 Gemma 4 12B 72%
  57. 57 Gemini 2.5 Pro May 2025 71.8%
  58. 58 Sarvam 105B 71.7%
  59. 59 North Mini Code 1.0 70.3%
  60. 60 Qwen3-VL-235B-A22B 70.1%
  61. 61 EXAONE 4.0 32B 70%
  62. 62 Sarvam 30B 70%
  63. 63 OpenReasoning-Nemotron-32B 69.8%
  64. 64 INTELLECT-3 69.3%
  65. 65 DiffusionGemma 26B A4B 69.1%
  66. 66 Qwen3-Next-80B-A3B 68.7%
  67. 67 Falcon H1R 7B 68.6%
  68. 68 Nemotron 3 Nano 30B-A3B 68.3%
  69. 69 NousCoder-14B 67.9%
  70. 70 o3-mini high 67.4%
  71. 71 seed-oss-36b-instruct 67.4%
  72. 72 Step 3 67.1%
  73. 73 OpenCodeReasoning-Nemotron-1.1-32B 66.8%
  74. 74 Grok-3 mini high 66.7%
  75. 75 Qwen3-235B-A22B 65.9%
  76. 76 Qwen3.5-9B 65.6%
  77. 77 Qwen3.5-Omni-Plus 65.6%
  78. 78 Ling-2.6-1T 65.6%
  79. 79 MiniMax-M1-80k 65%
  80. 80 Ministral 3 14B 64.6%
  81. 81 glm-4.7-flash 64%
  82. 82 Code World Model (CWM) 63.5%
  83. 83 Beam 1.0 Base 63.4%
  84. 84 Gemini 2.5 Flash May 2025 61.9%
  85. 85 Hermes 4 405B 61.4%
  86. 86 Gemini 2.5 Flash Apr 2025 60.6%
  87. 87 LFM2.5-2.6B 59.4%
  88. 88 Qwen3-Coder-Next 58.9%
  89. 89 Claude Opus 4 thinking 56.6%
  90. 90 MiniCPM-V-4.7-35B-A3B 56.6%
  91. 91 Claude Sonnet 4 thinking 55.9%
  92. 92 Qwen3.5-4B 55.8%
  93. 93 Kimi K2 Jul 2025 53.7%
  94. 94 LongCat-Flash-Omni 52.6%
  95. 95 Gemma 4 E4B 52%
  96. 96 Magistral Medium 1.0 50.3%
  97. 97 MiMo-7B-RL 49.3%
  98. 98 LongCat-Flash-Chat 48%
  99. 99 Magistral Small 1.0 47.4%
  100. 100 Reka Flash 3.1 46.8%
  101. 101 Gemma 4 E2B 44%
  102. 102 ERNIE-4.5-300B-A47B 38.8%
  103. 103 Claude 3.5 Sonnet (October 2024) 36.4%
  104. 104 Seed Diffusion Preview 33.7%
  105. 105 Mistral Medium 3 30.3%
  106. 106 GPT-4o Aug 2024 29.5%
  107. 107 GPT-4 Turbo Apr 2024 28.7%
  108. 108 GPT-4o mini 27.5%
  109. 109 DeepSeek-V3 27.2%
  110. 110 LFM2-8B-A1B 21%
  111. 111 Claude 3 Haiku 20.2%

Sources: Labs’ model cards and official LiveCodeBench v6 leaderboard

#ModelScoreSettingDateSource
1 DeepSeek-V4-Pro 93.5% Apr 2026 Lab model card
2 Qwen3.8-Omni-Flash 92.6% Sep 2026 Lab model card
3 Solar Open 2 92.4% Jul 2026 Lab model card
4 Qwen3.8 Flash 91.9% Aug 2026 Lab model card
5 DeepSeek-V4-Flash 91.6% Apr 2026 Lab model card
6 Qwen 3.8 27B 90.3% Aug 2026 Lab model card
7 Kimi K2.6 89.6% Apr 2026 Lab model card
8 Ring-2.5-1T 89.4% Feb 2026 Lab model card
9 Nemotron 3 Ultra 89% Jun 2026 Lab model card
10 Nemotron 3.5 Super VL 88.2% Oct 2026 Lab model card
11 Seed2.0 Pro 87.8% Feb 2026 Lab model card
12 MAI-Thinking-1 87.7% Jun 2026 Lab model card
13 Gemini 2.5 Deep Think 87.6% Aug 2025 Lab model card
14 Nemotron-Cascade 2 30B-A3B 87.2% Mar 2026 Lab model card
15 Qwen 3.6 Plus 87.1% Apr 2026 Lab model card
16 Ring-2.6-1T 87% Jun 2026 Lab model card
17 Step 3.5 Flash 86.4% Feb 2026 Lab model card
18 Qwen3-Max-2026-01-23 85.9% Jan 2026 Lab model card
19 MiMo-V2-Flash 85.1% Jan 2026 Lab model card
20 Kimi K2.5 85% Jan 2026 Lab model card
21 GLM-4.7 84.9% Dec 2025 Lab model card
22 Qwen3.6 27B 83.9% Apr 2026 Lab model card
23 Qwen3.5 397B-A17B 83.6% Feb 2026 Lab model card
24 DeepSeek-V3.2 83.3% Dec 2025 Lab model card
25 Kimi K2 Thinking 83.1% Nov 2025 Lab model card
26 GLM-4.6 82.8% Sep 2025 Lab model card
27 LongCat-Flash-Thinking-2601 82.8% Jan 2026 Lab model card
28 EXAONE 4.5 33B 81.4% Apr 2026 Lab model card
29 K-EXAONE 80.7% Dec 2025 Lab model card
30 Qwen3.5-27B 80.7% Feb 2026 Lab model card
31 Qwen 3.6 35B-A3B 80.4% Apr 2026 Lab model card
32 Flint 80.3% Oct 2026 Lab model card
33 Gemma 4 31B IT 80% Apr 2026 Lab model card
34 Seed1.8 79.5% Dec 2025 Lab model card
35 LongCat-Flash-Thinking 79.4% Sep 2025 Lab model card
36 Qwen3.5-122B-A10B 78.9% Feb 2026 Lab model card
37 Nemotron 3 Super 78.7% Mar 2026 Lab model card
38 Ring-1T 78.3% Oct 2025 Lab model card
39 Solar Pro 3 (March 2026 update) 77.7% Mar 2026 Lab model card
40 Gemma 4 26B A4B 77.1% Apr 2026 Lab model card
41 Keye-VL-2.0-30B-A3B 77.1% May 2026 Lab model card
42 ERNIE 5.0 76.2% Feb 2026 Lab model card
43 STEP3-VL-10B 75.8% Jan 2026 Lab model card
44 Granite 4.2 30B 75.8% Aug 2026 Lab model card
45 DeepSeek-V3.1 74.8% Aug 2025 Lab model card
46 Qwen3.5-35B-A3B 74.6% Feb 2026 Lab model card
47 Intern-S1-Pro 74.3% Feb 2026 Lab model card
48 Solar Open 100B 74.2% Jan 2026 Lab model card
49 Qwen3-235B-A22B-Thinking Jul 2025 74.1% Jul 2025 Lab model card
50 DeepSeek-V3.2-Exp 74.1% Sep 2025 Lab model card
51 LFM3-1.7B 73.5% Oct 2026 Lab model card
52 DeepSeek-R1 May 2025 73.3% May 2025 Lab model card
53 Gemma 4 12B 72% Jun 2026 Lab model card
54 Sarvam 105B 71.7% Mar 2026 Lab model card
55 North Mini Code 1.0 70.3% Jun 2026 Lab model card
56 Qwen3-VL-235B-A22B 70.1% Sep 2025 Lab model card
57 Sarvam 30B 70% Mar 2026 Lab model card
58 INTELLECT-3 69.3% Dec 2025 Lab model card
59 DiffusionGemma 26B A4B 69.1% Jun 2026 Lab model card
60 Qwen3-Next-80B-A3B 68.7% Sep 2025 Lab model card
61 Falcon H1R 7B 68.6% Jan 2026 Lab model card
62 Nemotron 3 Nano 30B-A3B 68.3% Dec 2025 Lab model card
63 NousCoder-14B 67.9% Jan 2026 Lab model card
64 seed-oss-36b-instruct 67.4% Aug 2025 Lab model card
65 Step 3 67.1% Jul 2025 Lab model card
66 EXAONE 4.0 32B 66.7% Jul 2025 Lab model card
67 Qwen3.5-9B 65.6% Mar 2026 Lab model card
68 Qwen3.5-Omni-Plus 65.6% Mar 2026 Lab model card
69 Ling-2.6-1T 65.6% Jun 2026 Lab model card
70 MiniMax-M1-80k 65% Jun 2025 Lab model card
71 Ministral 3 14B 64.6% Jan 2026 Lab model card
72 glm-4.7-flash 64% Jan 2026 Lab model card
73 Code World Model (CWM) 63.5% Sep 2025 Lab model card
74 Beam 1.0 Base 63.4% Oct 2026 Lab model card
75 Hermes 4 405B 61.4% Aug 2025 Lab model card
76 LFM2.5-2.6B 59.4% Aug 2026 Lab model card
77 Qwen3-Coder-Next 58.9% Feb 2026 Lab model card
78 MiniCPM-V-4.7-35B-A3B 56.6% Oct 2026 Lab model card
79 Qwen3.5-4B 55.8% Mar 2026 Lab model card
80 Kimi K2 Jul 2025 53.7% Jul 2025 Lab model card
81 LongCat-Flash-Omni 52.6% Oct 2025 Lab model card
82 Gemma 4 E4B 52% Apr 2026 Lab model card
83 Magistral Medium 1.0 50.3% Jun 2025 Lab model card
84 MiMo-7B-RL 49.3% Apr 2025 Lab model card
85 LongCat-Flash-Chat 48% Sep 2025 Lab model card
86 Magistral Small 1.0 47.4% Jun 2025 Lab model card
87 Reka Flash 3.1 46.8% May 2025 Lab model card
88 Gemma 4 E2B 44% Apr 2026 Lab model card
89 ERNIE-4.5-300B-A47B 38.8% Jun 2025 Lab model card
90 Seed Diffusion Preview 33.7% Jul 2025 Lab model card
91 Mistral Medium 3 30.3% May 2025 Lab model card
92 LFM2-8B-A1B 21% Oct 2025 Lab model card
#ModelScoreSettingDateSource
1 o4-mini 80.2% high Apr 2025 Official leaderboard
2 o3 75.8% high Apr 2025 Official leaderboard
3 Gemini 2.5 Pro Jun 2025 73.6% Jun 2025 Official leaderboard
4 DeepSeek-R1 May 2025 73.1% May 2025 Official leaderboard
5 Gemini 2.5 Pro May 2025 71.8% May 2025 Official leaderboard
6 EXAONE 4.0 32B 70% Jul 2025 Official leaderboard
7 OpenReasoning-Nemotron-32B 69.8% Jul 2025 Official leaderboard
8 o3-mini 67.4% high Jan 2025 Official leaderboard
9 OpenCodeReasoning-Nemotron-1.1-32B 66.8% Jun 2025 Official leaderboard
10 Grok-3 mini 66.7% high Apr 2025 Official leaderboard
11 Qwen3-235B-A22B 65.9% Apr 2025 Official leaderboard
12 Gemini 2.5 Flash May 2025 61.9% May 2025 Official leaderboard
13 Gemini 2.5 Flash Apr 2025 60.6% Apr 2025 Official leaderboard
14 Claude Opus 4 56.6% thinking May 2025 Official leaderboard
15 Claude Sonnet 4 55.9% thinking May 2025 Official leaderboard
16 Claude 3.5 Sonnet (October 2024) 36.4% Oct 2024 Official leaderboard
17 GPT-4o Aug 2024 29.5% Aug 2024 Official leaderboard
18 GPT-4 Turbo Apr 2024 28.7% Apr 2024 Official leaderboard
19 GPT-4o mini 27.5% Jul 2024 Official leaderboard
20 DeepSeek-V3 27.2% Dec 2024 Official leaderboard
21 Claude 3 Haiku 20.2% Mar 2024 Official leaderboard
#ModelScoreSettingDateSource
1 DeepSeek-V4-Pro 93.5% Apr 2026 Lab model card
2 Qwen3.8-Omni-Flash 92.6% Sep 2026 Lab model card
3 Solar Open 2 92.4% Jul 2026 Lab model card
4 Qwen3.8 Flash 91.9% Aug 2026 Lab model card
5 DeepSeek-V4-Flash 91.6% Apr 2026 Lab model card
6 Qwen 3.8 27B 90.3% Aug 2026 Lab model card
7 Kimi K2.6 89.6% Apr 2026 Lab model card
8 Ring-2.5-1T 89.4% Feb 2026 Lab model card
9 Nemotron 3 Ultra 89% Jun 2026 Lab model card
10 Nemotron 3.5 Super VL 88.2% Oct 2026 Lab model card
11 Seed2.0 Pro 87.8% Feb 2026 Lab model card
12 MAI-Thinking-1 87.7% Jun 2026 Lab model card
13 Gemini 2.5 Deep Think 87.6% Aug 2025 Lab model card
14 Nemotron-Cascade 2 30B-A3B 87.2% Mar 2026 Lab model card
15 Qwen 3.6 Plus 87.1% Apr 2026 Lab model card
16 Ring-2.6-1T 87% Jun 2026 Lab model card
17 Step 3.5 Flash 86.4% Feb 2026 Lab model card
18 Qwen3-Max-2026-01-23 85.9% Jan 2026 Lab model card
19 MiMo-V2-Flash 85.1% Jan 2026 Lab model card
20 Kimi K2.5 85% Jan 2026 Lab model card
21 GLM-4.7 84.9% Dec 2025 Lab model card
22 Qwen3.6 27B 83.9% Apr 2026 Lab model card
23 Qwen3.5 397B-A17B 83.6% Feb 2026 Lab model card
24 DeepSeek-V3.2 83.3% Dec 2025 Lab model card
25 Kimi K2 Thinking 83.1% Nov 2025 Lab model card
26 GLM-4.6 82.8% Sep 2025 Lab model card
27 LongCat-Flash-Thinking-2601 82.8% Jan 2026 Lab model card
28 EXAONE 4.5 33B 81.4% Apr 2026 Lab model card
29 K-EXAONE 80.7% Dec 2025 Lab model card
30 Qwen3.5-27B 80.7% Feb 2026 Lab model card
31 Qwen 3.6 35B-A3B 80.4% Apr 2026 Lab model card
32 Flint 80.3% Oct 2026 Lab model card
33 o4-mini 80.2% high Apr 2025 Official leaderboard
34 Gemma 4 31B IT 80% Apr 2026 Lab model card
35 Seed1.8 79.5% Dec 2025 Lab model card
36 LongCat-Flash-Thinking 79.4% Sep 2025 Lab model card
37 Qwen3.5-122B-A10B 78.9% Feb 2026 Lab model card
38 Nemotron 3 Super 78.7% Mar 2026 Lab model card
39 Ring-1T 78.3% Oct 2025 Lab model card
40 Solar Pro 3 (March 2026 update) 77.7% Mar 2026 Lab model card
41 Gemma 4 26B A4B 77.1% Apr 2026 Lab model card
42 Keye-VL-2.0-30B-A3B 77.1% May 2026 Lab model card
43 ERNIE 5.0 76.2% Feb 2026 Lab model card
44 o3 75.8% high Apr 2025 Official leaderboard
45 STEP3-VL-10B 75.8% Jan 2026 Lab model card
46 Granite 4.2 30B 75.8% Aug 2026 Lab model card
47 DeepSeek-V3.1 74.8% Aug 2025 Lab model card
48 Qwen3.5-35B-A3B 74.6% Feb 2026 Lab model card
49 Intern-S1-Pro 74.3% Feb 2026 Lab model card
50 Solar Open 100B 74.2% Jan 2026 Lab model card
51 Qwen3-235B-A22B-Thinking Jul 2025 74.1% Jul 2025 Lab model card
52 DeepSeek-V3.2-Exp 74.1% Sep 2025 Lab model card
53 Gemini 2.5 Pro Jun 2025 73.6% Jun 2025 Official leaderboard
54 LFM3-1.7B 73.5% Oct 2026 Lab model card
55 DeepSeek-R1 May 2025 73.3% May 2025 Lab model card
56 Gemma 4 12B 72% Jun 2026 Lab model card
57 Gemini 2.5 Pro May 2025 71.8% May 2025 Official leaderboard
58 Sarvam 105B 71.7% Mar 2026 Lab model card
59 North Mini Code 1.0 70.3% Jun 2026 Lab model card
60 Qwen3-VL-235B-A22B 70.1% Sep 2025 Lab model card
61 EXAONE 4.0 32B 70% Jul 2025 Official leaderboard
62 Sarvam 30B 70% Mar 2026 Lab model card
63 OpenReasoning-Nemotron-32B 69.8% Jul 2025 Official leaderboard
64 INTELLECT-3 69.3% Dec 2025 Lab model card
65 DiffusionGemma 26B A4B 69.1% Jun 2026 Lab model card
66 Qwen3-Next-80B-A3B 68.7% Sep 2025 Lab model card
67 Falcon H1R 7B 68.6% Jan 2026 Lab model card
68 Nemotron 3 Nano 30B-A3B 68.3% Dec 2025 Lab model card
69 NousCoder-14B 67.9% Jan 2026 Lab model card
70 o3-mini 67.4% high Jan 2025 Official leaderboard
71 seed-oss-36b-instruct 67.4% Aug 2025 Lab model card
72 Step 3 67.1% Jul 2025 Lab model card
73 OpenCodeReasoning-Nemotron-1.1-32B 66.8% Jun 2025 Official leaderboard
74 Grok-3 mini 66.7% high Apr 2025 Official leaderboard
75 Qwen3-235B-A22B 65.9% Apr 2025 Official leaderboard
76 Qwen3.5-9B 65.6% Mar 2026 Lab model card
77 Qwen3.5-Omni-Plus 65.6% Mar 2026 Lab model card
78 Ling-2.6-1T 65.6% Jun 2026 Lab model card
79 MiniMax-M1-80k 65% Jun 2025 Lab model card
80 Ministral 3 14B 64.6% Jan 2026 Lab model card
81 glm-4.7-flash 64% Jan 2026 Lab model card
82 Code World Model (CWM) 63.5% Sep 2025 Lab model card
83 Beam 1.0 Base 63.4% Oct 2026 Lab model card
84 Gemini 2.5 Flash May 2025 61.9% May 2025 Official leaderboard
85 Hermes 4 405B 61.4% Aug 2025 Lab model card
86 Gemini 2.5 Flash Apr 2025 60.6% Apr 2025 Official leaderboard
87 LFM2.5-2.6B 59.4% Aug 2026 Lab model card
88 Qwen3-Coder-Next 58.9% Feb 2026 Lab model card
89 Claude Opus 4 56.6% thinking May 2025 Official leaderboard
90 MiniCPM-V-4.7-35B-A3B 56.6% Oct 2026 Lab model card
91 Claude Sonnet 4 55.9% thinking May 2025 Official leaderboard
92 Qwen3.5-4B 55.8% Mar 2026 Lab model card
93 Kimi K2 Jul 2025 53.7% Jul 2025 Lab model card
94 LongCat-Flash-Omni 52.6% Oct 2025 Lab model card
95 Gemma 4 E4B 52% Apr 2026 Lab model card
96 Magistral Medium 1.0 50.3% Jun 2025 Lab model card
97 MiMo-7B-RL 49.3% Apr 2025 Lab model card
98 LongCat-Flash-Chat 48% Sep 2025 Lab model card
99 Magistral Small 1.0 47.4% Jun 2025 Lab model card
100 Reka Flash 3.1 46.8% May 2025 Lab model card
101 Gemma 4 E2B 44% Apr 2026 Lab model card
102 ERNIE-4.5-300B-A47B 38.8% Jun 2025 Lab model card
103 Claude 3.5 Sonnet (October 2024) 36.4% Oct 2024 Official leaderboard
104 Seed Diffusion Preview 33.7% Jul 2025 Lab model card
105 Mistral Medium 3 30.3% May 2025 Lab model card
106 GPT-4o Aug 2024 29.5% Aug 2024 Official leaderboard
107 GPT-4 Turbo Apr 2024 28.7% Apr 2024 Official leaderboard
108 GPT-4o mini 27.5% Jul 2024 Official leaderboard
109 DeepSeek-V3 27.2% Dec 2024 Official leaderboard
110 LFM2-8B-A1B 21% Oct 2025 Lab model card
111 Claude 3 Haiku 20.2% Mar 2024 Official leaderboard

The benchmark's own leaderboard has 22 entries, 21 of them models we track. Switch the scores above to rank them together.

More effort adds a median 14 points

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
o4-mini
low medium high
80.2%
o3-mini
low medium high
67.4%
Claude Opus 4
thinking
56.6%
Claude Sonnet 4
thinking
55.9%

Source: Official LiveCodeBench v6 leaderboard

8 models have held number one in labs’ own numbers since its release in April 2025 2 models have held number one across all independent runs since its release in April 2025 7 models have held number one with claims included since its release in April 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 Xiaomi Corp DeepSeek Qwen Google DeepMind Ant Group Moonshot AI
Own leaderboard Not tested since Jul 2025
o4-mini 80.2%
Lab claims Passed 85% in Aug 2025, by Gemini 2.5 Deep Think

Sources: Official LiveCodeBench v6 leaderboard and labs’ model cards

Google DeepMind’s own number for Gemini 2.5 Deep Think passed 85% in August 2025

That was 4 months after its release. By the same age, 13% of the benchmarks from 2025 had passed 85%. 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% 2024 2025 2026 2027 Near the ceiling (85%) DeepSeek-V4-Pro Gemini 2.5 Deep Think o4-mini o3-mini Claude 3.5 Sonnet (October 2024) GPT-4o Claude 3 Haiku 0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Near the ceiling (85%) DeepSeek-V4-Pro o4-mini

Sources: Official LiveCodeBench v6 leaderboard and labs’ model cards

10 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.

Xiaomi Corp
Mistral AI
DeepSeek
Reka
MiniMax
Moonshot AI
LG AI Research
Qwen
StepFun
ByteDance Seed
Google DeepMind
Nous Research
Meta
Z.ai
Liquid AI
Ant Group
NVIDIA
Upstage
Technology Innovation Institute
Meituan
Shanghai AI Lab
Baidu
Sarvam AI
Kuaishou
Microsoft
Cohere
IBM
OpenBMB (ModelBest)
Recursive Superintelligence
DateLaunch postScoreAs the post writes it
Oct 27, 2026 LFM3-1.7B 73.5 LiveCodeBench v6
Oct 21, 2026 Flint 80.3 LiveCodeBench v6
Oct 15, 2026 Nemotron 3.5 Super VL 88.2 LiveCodeBench v6
Oct 7, 2026 MiniCPM-V-4.7-35B-A3B 56.6 LiveCodeBench-v6
Sep 18, 2026 Qwen3.8-Omni-Flash 92.6 LiveCodeBench v6
Aug 26, 2026 Qwen3.8 Flash 91.9 Competitive coding LiveCodeBench v6
Aug 25, 2026 Granite 4.2 30B 75.8 LiveCodeBench v6
Aug 14, 2026 Qwen 3.8 27B 90.3 Competitive coding LiveCodeBench v6
Aug 4, 2026 LFM2.5-2.6B 59.4 LiveCodeBenchv6
Jul 22, 2026 Solar Open 2 92.4 LiveCodeBench (v6)
Jun 10, 2026 DiffusionGemma 26B A4B 69.1 LiveCodeBench v6
Jun 9, 2026 North Mini Code 1.0 70.3 LiveCodeBench v6
Jun 4, 2026 Nemotron 3 Ultra 89 LiveCodeBench (v6)
Jun 3, 2026 Gemma 4 12B 72 LiveCodeBench v6
Jun 2, 2026 MAI-Thinking-1 87.7 LCB v6
May 25, 2026 Keye-VL-2.0-30B-A3B 77.1 LiveCodeBench v6
Apr 24, 2026 DeepSeek-V4-Flash 91.6 LiveCodeBench (Pass@1), V4-Flash Max
Apr 22, 2026 Qwen3.6 27B 83.9 LiveCodeBench v6
Apr 20, 2026 Kimi K2.6 89.6 LiveCodeBench (v6)
Apr 15, 2026 Qwen 3.6 35B-A3B 80.4 LiveCodeBench v6
Apr 9, 2026 EXAONE 4.5 33B 81.4 LiveCodeBench v6
Apr 2, 2026 Gemma 4 E4B 52 LiveCodeBench v6
Apr 2, 2026 Gemma 4 E2B 44 LiveCodeBench v6
Apr 2, 2026 Gemma 4 26B A4B 77.1 LiveCodeBench v6
Apr 2, 2026 Qwen 3.6 Plus 87.1 LiveCodeBench v6
Apr 2, 2026 Gemma 4 31B IT 80 LiveCodeBench v6
Mar 30, 2026 Qwen3.5-Omni-Plus 65.6 LiveCodeBench v6
Mar 24, 2026 Solar Pro 3 (March 2026 update) 77.7 LCB.v1-6 (cumulative)
Mar 19, 2026 Nemotron-Cascade 2 30B-A3B 87.2 LiveCodeBench v6 (2408-2505) (TIR 88.4)
Mar 6, 2026 Sarvam 30B 70 Live Code Bench v6
Mar 6, 2026 Sarvam 105B 71.7 Live Code Bench v6
Mar 2, 2026 Qwen3.5-4B 55.8 LiveCodeBench v6
Mar 2, 2026 Qwen3.5-9B 65.6 LiveCodeBench v6
Feb 24, 2026 Qwen3.5-35B-A3B 74.6 LiveCodeBench v6
Feb 24, 2026 Qwen3.5-27B 80.7 LiveCodeBench v6
Feb 24, 2026 Qwen3.5-122B-A10B 78.9 LiveCodeBench v6
Feb 16, 2026 Qwen3.5 397B-A17B 83.6 LiveCodeBench v6
Feb 14, 2026 Seed2.0 Pro 87.8 LiveCodeBench (v6)
Feb 10, 2026 Ring-2.5-1T 89.4 Best of LiveCodeBench-v6 (Mean@4) 87.89 / Heavy Thinking 89.40
Feb 6, 2026 ERNIE 5.0 76.2 LiveCodeBench v6
Feb 2, 2026 Intern-S1-Pro 74.3 LCB V6
Feb 2, 2026 Step 3.5 Flash 86.4 LiveCodeBench-V6
Jan 27, 2026 Kimi K2.5 85 LiveCodeBench (v6)
Jan 23, 2026 Qwen3-Max-2026-01-23 85.9 LiveCodeBench v6
Jan 19, 2026 glm-4.7-flash 64 LCB v6
Jan 14, 2026 LongCat-Flash-Thinking-2601 82.8 LCB (24.08-25.05) (Avg@4)
Jan 6, 2026 NousCoder-14B 67.9 LiveCodeBench v6 (08/01/2024 - 05/01/2025), Pass@1
Jan 5, 2026 Falcon H1R 7B 68.6 LCB v6
Jan 5, 2026 Solar Open 100B 74.2 LiveCodeBench (v1–v6 cumul)
Dec 31, 2025 K-EXAONE 80.7 LiveCodeBench v6
Dec 22, 2025 GLM-4.7 84.9 LiveCodeBench-v6
Dec 18, 2025 Seed1.8 79.5 LiveCodeBench(v6)
Dec 16, 2025 MiMo-V2-Flash 80.6 LiveCodeBench-v6
Dec 15, 2025 Nemotron 3 Nano 30B-A3B 68.3 LiveCodeBench (v6 2025-08–2025-05)
Nov 6, 2025 Kimi K2 Thinking 83.1 LiveCodeBenchV6, no tools
Oct 10, 2025 Ring-1T 78.3 LiveCodeBench(2408-2505) (pass@1)
Oct 7, 2025 LFM2-8B-A1B 21 LCB v6
Sep 30, 2025 GLM-4.6 82.8 LiveCodeBench v6
Sep 24, 2025 Code World Model (CWM) 63.5 LCBv6
Sep 24, 2025 Qwen3-Max LiveCodeBench v6 (25.02-25.05)
Sep 23, 2025 Qwen3-VL-235B-A22B 70.1 LCB v6
Sep 11, 2025 Qwen3-Next-80B-A3B 68.7 LiveCodeBench v6 (25.02-25.05)
Aug 26, 2025 Hermes 4 405B 61.3 LCBv6 Aug2024+ (reasoning)
Aug 21, 2025 DeepSeek-V3.1 74.8 LiveCodeBench (2408-2505) (Pass@1)
Aug 20, 2025 seed-oss-36b-instruct 67.4 LiveCodeBench v6 (02/2025-05/2025)
Aug 1, 2025 Gemini 2.5 Deep Think 87.6 LiveCodeBench v6 (UI: 1/1/2025-5/1/2025)
Jul 31, 2025 Seed Diffusion Preview 33.7 LiveCodeBench* (1055 problems from v1 to v6, Figure 1)
Jul 31, 2025 Step 3 67.1 LiveCodeBench (202408-202505)
Jul 25, 2025 Qwen3-235B-A22B-Thinking 74.1 LiveCodeBench v6 (25.02-25.05)
Jul 15, 2025 EXAONE 4.0 32B 66.7 LiveCodeBench v6 (reasoning)
Jul 11, 2025 Kimi K2 53.7 LiveCodeBench v6 (Aug 24 - May 25), Pass@1
Jun 16, 2025 MiniMax-M1-80k 65 LiveCodeBench (24/8~25/5)
May 28, 2025 Reka Flash 3.1 46.8 LCB_v6 (25.2-25.5)
May 28, 2025 DeepSeek-R1 73.3 LiveCodeBench (2408-2505) (Pass@1)
May 7, 2025 Mistral Medium 3 30.3 LiveCodeBench (v6) 0-shot
Apr 29, 2025 MiMo-7B-RL 49.3 LiveCodeBench v6 (Pass@1, 20250201-20250501)

53 system cards report it, from 23 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 15, 2026 Nemotron 3.5 Super VL technical report 88.2 LiveCodeBench v6, 454 problems Aug 2024 to May 2025 (Table 10) Capabilities
Oct 5, 2026 Beam 1.0 Base technical report 63.4 LCB v6, 0-shot greedy, Table 23 Capabilities
Sep 22, 2026 Qwen3.8-Omni-Flash technical report 92.6 LiveCodeBench v6 (Table 2) Capabilities
Jul 31, 2026 DiffusionGemma 26B A4B technical report 69.1 LiveCodeBench-V6, TD, 60.6 no-think, 71.4 AR (Table 3) Capabilities
Jul 24, 2026 Gemma 4 E4B technical report 52 LiveCodeBench v6, thinking (Table 5) Capabilities
Jul 24, 2026 Gemma 4 E2B technical report 44 LiveCodeBench v6, thinking (Table 5) Capabilities
Jul 24, 2026 Gemma 4 31B IT technical report 80 LiveCodeBench v6, thinking (Table 5) Capabilities
Jul 24, 2026 Gemma 4 26B A4B technical report 77.1 LiveCodeBench v6, thinking (Table 5) Capabilities
Jul 22, 2026 Solar Open 2 technical report 92.4 LiveCodeBench (v6) (Table 3) Capabilities
Jun 13, 2026 Ring-2.6-1T technical report 87 LCB-v6 (Avg@4), 2408-2505, xhigh (Table 6). High gives 76.71 Capabilities
Jun 13, 2026 Ling-2.6-1T technical report 65.6 LiveCodeBench-v6, Mean@4 (Table 5) Capabilities
Jun 9, 2026 Nemotron 3 Ultra technical report 89 LiveCodeBench (v6) (Table 10) Capabilities
Jun 9, 2026 Keye-VL-2.0-30B-A3B technical report 64.2 LiveCodeBench v6 (Table 3 and Section 6.2.1 text). Figure 5 prints 77.1 for the same row Capabilities
Jun 2, 2026 MAI-Thinking-1 technical report 87.7 LCB v6, pass@1, 256k max output tokens (Table 11) Capabilities
Apr 26, 2026 DeepSeek-V4-Pro technical report 93.5 LiveCodeBench (Pass@1), LiveCodeBench-v6 per 5.3.1, DS-V4-Pro Max (Table 6) Capabilities
Apr 26, 2026 DeepSeek-V4-Flash technical report 91.6 LiveCodeBench (Pass@1-COT), LiveCodeBench-v6 per 5.3.1, DeepSeek-V4-Flash Max (Table 7) Capabilities
Apr 17, 2026 Qwen3.5-Omni-Plus technical report 65.6 LiveCodeBench v6 (Table 4) Capabilities
Apr 9, 2026 EXAONE 4.5 33B technical report 81.4 LiveCodeBench v6 (Table 3, Reasoning mode) Capabilities
Mar 26, 2026 Intern-S1-Pro technical report 74.3 LCB V6, thinking (Table 2) Capabilities
Mar 21, 2026 Seed1.8 model card 79.5 LiveCodeBench(v6) (Table 1). Table 7 prints 79.4 Avg@8 and 84.7 with increased thinking Capabilities
Mar 19, 2026 Nemotron-Cascade 2 30B-A3B technical report 87.2 LiveCodeBench v6 (2408-2505), avg@8 (Tables 1 and 6). TIR 88.4 Capabilities
Mar 11, 2026 Nemotron 3 Super technical report 78.7 LiveCodeBench (v6 2024-08 to 2025-05), BF16 column (Table 8) Capabilities
Feb 28, 2026 Qwen3-Coder-Next technical report 58.9 LiveCodeBench (v6) (Table 6) Capabilities
Feb 14, 2026 Seed2.0 Pro model card 87.8 LiveCodeBench (v6) (Table 3) Capabilities
Feb 11, 2026 Step 3.5 Flash technical report 86.4 LiveCodeBench-v6 (2024.08 to 2025.05) (Table 5). PaCoRe 88.9 Capabilities
Feb 4, 2026 ERNIE 5.0 technical report 76.2 LiveCodeBench v6 (2408 to 2505) (Table 2) Capabilities
Jan 30, 2026 Kimi K2.5 technical report 85 LiveCodeBench (v6) (Table 4) Capabilities
Jan 23, 2026 LongCat-Flash-Thinking-2601 technical report 82.8 LCB (24.08-25.05) (Avg@4), 454 problems (Table 2) Capabilities
Jan 14, 2026 STEP3-VL-10B technical report 75.8 LiveCodeBench (2408-2505), the v6 window, SeRe (Tables 2 and 3). PaCoRe 76.43 Capabilities
Jan 13, 2026 Ministral 3 14B technical report 64.6 LiveCodeBench v6, pass@5 (Table 5) Capabilities
Jan 11, 2026 Solar Open 100B technical report 74.2 LiveCodeBench (v1 to v6 cumul) (Table 6), called LiveCodeBench v6 in 7.1 Capabilities
Jan 6, 2026 MiMo-V2-Flash technical report 85.1 LiveCodeBench-v6 (Table 9). The setup text names LiveCodeBench 2024.08 to 2025.04, Table 7 gives 83.2, and the launch model card prints 80.6 Capabilities
Jan 5, 2026 Falcon H1R 7B technical report 68.6 LCB v6 (Table 5) Capabilities
Jan 5, 2026 K-EXAONE technical report 80.7 LiveCodeBench v6 (Table 3, Reasoning mode). Non-reasoning 44.6 (Table 4) Capabilities
Dec 23, 2025 Nemotron 3 Nano 30B-A3B technical report 68.3 LiveCodeBench (v6 2024-08 to 2025-05) (Table 3) Capabilities
Dec 18, 2025 INTELLECT-3 technical report 69.3 LCB v6 (2408-2505, 454 problems), Avg@2 (Table 2, Appendix A.1) Capabilities
Dec 1, 2025 DeepSeek-V3.2 technical report 83.3 LiveCodeBench (2024.08-2025.04) (Pass@1-COT), DeepSeek-V3.2 Thinking (Table 2) Capabilities
Nov 26, 2025 Qwen3-VL-235B-A22B technical report 70.1 LiveCodeBench v6 (Table 6) Capabilities
Oct 31, 2025 LongCat-Flash-Omni technical report 52.6 LiveCodeBench (pass@1), problems 2024.08 to 2025.05 (Table 13) Capabilities
Oct 21, 2025 Ring-1T technical report 78.3 LCB-v6 (2408-2505) (Avg@4) (Table 1) Capabilities
Sep 30, 2025 Code World Model (CWM) technical report 63.5 LCBv6, 01.08.2024-01.05.2025, pass@1 (Table 12) Capabilities
Sep 29, 2025 DeepSeek-V3.2-Exp technical report 74.1 LiveCodeBench (2408-2505) (Pass@1), DeepSeek-V3.2-Exp (Table 1) Capabilities
Sep 23, 2025 LongCat-Flash-Thinking technical report 79.4 LCB (24.08-25.05)(Mean@4) (Table 2) Capabilities
Sep 1, 2025 LongCat-Flash-Chat technical report 48 LiveCodeBench (pass@1), problems 2024.08 to 2025.05 (Table 3) Capabilities
Aug 25, 2025 Hermes 4 405B technical report 61.4 LCBv6 Aug2024+, reasoning mode (Table 3). Non-reasoning 28.1. v1 printed 61.3 Capabilities
Aug 4, 2025 Seed Diffusion Preview technical report 33.7 LiveCodeBench*, 1,055 problems from v1 to v6 (Figure 1). The v6 result named in 4.1 is not printed Capabilities
Jul 21, 2025 Kimi K2 technical report 53.7 LiveCodeBench v6 (Pass@1), Aug 2024 to May 2025 (Table 3) Capabilities
Jul 15, 2025 EXAONE 4.0 32B technical report 66.7 LiveCodeBench v6, n=4, 64K budget (Table 3, Reasoning). 32K budget 67.3 (Table 7). Non-reasoning 43.1 (Table 4) Capabilities
Jun 29, 2025 ERNIE-4.5-300B-A47B technical report 38.8 LiveCodeBench (Table 5), named LiveCodeBench (v6, 20240801-20250501) in 8.1.2 Capabilities
Jun 16, 2025 MiniMax-M1-80k technical report 65 LiveCodeBench (24/8∼25/5), avg of 16 samples (Table 2) Capabilities
Jun 12, 2025 Magistral Small 1.0 technical report 47.4 LiveCodeBench (v6), SFT + RL (Magistral Small) column (Table 3) Capabilities
Jun 12, 2025 Magistral Medium 1.0 technical report 50.3 LiveCodeBench (v6) (Table 2) Capabilities
May 12, 2025 MiMo-7B-RL technical report 49.3 LiveCodeBench v6 (Pass@1), 20250201-20250501 (Table 4, Table 5) Capabilities

The lab's number comes out ahead in 1 of 2 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
EXAONE 4.0 32B 66.7% LiveCodeBench v6 (reasoning) 70% Leaderboard −3.3 pts
DeepSeek-R1 May 2025 73.3% LiveCodeBench (2408-2505) (Pass@1) 73.1% Leaderboard +0.2 pts

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.

DeepSeek Qwen Other labs Meituan MiniMax Ai2 Google DeepMind LG AI Research NVIDIA Mistral AI Xiaomi Corp Ant Group Moonshot AI
LiveCodeBench v6 Passed 85% in Aug 2025, by Gemini 2.5 Deep Think
LiveCodeBench Passed 85% in Dec 2025, by DeepSeek-V3.2-Speciale
LiveCodeBench v5 Not tested since Apr 2026
LiveCodeBench v6 Passed 85% in Aug 2025, by Gemini 2.5 Deep Think

Source: Labs’ model cards

More in Coding

All 523 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471
  • NL2Repo-Bench In use From one natural-language requirements document and an empty workspace, agent builds an installable Python library. Scored by original project's test pass rate. DeepSeek V4.1 Flash 1017 68.2%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (22 entries), launch posts (75 scores) and system cards (53 scores).
  • Pass Models scored 111 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from October 2026, for LFM3-1.7B.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Qwen 3.8 27B (#59).
  • Note Headroom Saturated. The best is 93.5%, past 85% since August 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on livecodebench.github.io: 27 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 19th, 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