Benchmark Almanac

Big Code Models Leaderboard

Qwen2.5-Coder-32B-Instruct 59.2% leads since November 2024
Not in model cards

Models generate programs in multiple languages. The board compares pass rates and averages each model’s win rate across languages.

The source CSV includes individual language scores and the average score.

Made by
BigCode
Released
July 2023
Kind
Function-level problems
Level
University
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Code
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Qwen2.5-Coder-32B #188 on the capability index
Last model added
November 2024
Scores refreshed
November 20th, 2026
60
Entries on its own leaderboard, 9 models we track

Qwen2.5-Coder-32B-Instruct leads with 59.2% among models released by November 2024 Qwen2.5-Coder-32B-Instruct leads with 59.2% among models released by November 2024

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in November 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

60 entries. Scroll sideways for the rest.

  1. 59.2 Qwen2.5-Coder-32B-Instruct
  2. 56.7 Qwen2.5-Coder-32B
  3. 56.3 OpenCodeInterpreter-DS-33B
  4. 55.9 Nxcode-CQ-7B-orpo
  5. 55.7 CodeQwen1.5-7B-Chat
  6. 54.7 CodeFuse-DeepSeek-33b
  7. 52.3 DeepSeek-Coder-33b-instruct
  8. 51.7 Artigenz-Coder-DS-6.7B
  9. 50.6 DeepSeek-Coder-7b-instruct
  10. 49.8 OpenCodeInterpreter-DS-6.7B
  11. 49.4 Phind-CodeLlama-34B-v2
  12. 48.1 Phind-CodeLlama-34B-v1
  13. 46.6 Phind-CodeLlama-34B-Python-v1
  14. 45.2 CodeQwen1.5-7B
  15. 43.6 WizardCoder-Python-34B-V1.0
  16. 43.6 Code Llama 70B Instruct
  17. 43.2 CodeLlama-70b
  18. 42.8 DeepSeek-Coder-33b-base
  19. 42.3 CodeLlama-70b-Python
  20. 39.9 StarCoder 2 15B
  21. 38.6 Code Llama 34B Instruct
  22. 38.4 WizardCoder-Python-13B-V1.0
  23. 38.3 DeepSeek-Coder-7b-base
  24. 37.8 CodeLlama-34b
  25. 37.0 CodeLlama-34b-Python
  26. 35.3 WizardCoder-15B-V1.0
  27. 34.4 CodeLlama-13b-Instruct
  28. 33.8 CodeGemma-7B
  29. 32.3 CodeLlama-13b
  30. 30.2 CodeLlama-13b-Python
  31. 28.5 CodeLlama-7b-Instruct
  32. 28.3 StarCoder 2 7B
  33. 27.0 CodeGemma-7B-it
  34. 26.9 CodeShell-7B
  35. 26.9 CodeLlama-7b
  36. 25.3 OctoCoder-15B
  37. 25.1 CodeLlama-7b-Python
  38. 24.6 StarCoder-15B
  39. 24.2 Falcon-180B
  40. 24.0 StarCoderBase-15B
  41. 22.8 StarCoder 2 3B
  42. 20.6 CodeGeex2-6B
  43. 20.1 StarCoderBase-7B
  44. 19.9 OctoGeeX-7B
  45. 18.6 WizardCoder-3B-V1.0
  46. 18.2 CodeGen25-7B-multi
  47. 17.5 CodeGemma-2B
  48. 17.2 Refact-1.6B
  49. 16.2 Stable-code-3b
  50. 16.0 DeepSeek-Coder-1b-base
  51. 13.9 StarCoderBase-3B
  52. 12.4 WizardCoder-1B-V1.0
  53. 10.5 Replit-2.7B
  54. 10.1 CodeGen25-7B-mono
  55. 10.0 StarCoderBase-1.1B
  56. 8.92 CodeGen-16B-Multi
  57. 7.88 StableCode-3B-alpha
  58. 7.65 DeciCoder-1B
  59. 6.92 Phi-1
  60. 6.27 SantaCoder-1.1B
  1. 1 Qwen2.5-Coder-32B-Instruct 59.2%
  2. 2 Qwen2.5-Coder-32B 56.7%
  3. 3 OpenCodeInterpreter-DS-33B 56.3%
  4. 4 Nxcode-CQ-7B-orpo 55.9%
  5. 5 CodeQwen1.5-7B-Chat 55.7%
  6. 6 CodeFuse-DeepSeek-33b 54.7%
  7. 7 DeepSeek-Coder-33b-instruct 52.3%
  8. 8 Artigenz-Coder-DS-6.7B 51.7%
  9. 9 DeepSeek-Coder-7b-instruct 50.6%
  10. 10 OpenCodeInterpreter-DS-6.7B 49.8%
  11. 11 Phind-CodeLlama-34B-v2 49.4%
  12. 12 Phind-CodeLlama-34B-v1 48.1%
  13. 13 Phind-CodeLlama-34B-Python-v1 46.6%
  14. 14 CodeQwen1.5-7B 45.2%
  15. 15 WizardCoder-Python-34B-V1.0 43.6%
  16. 16 Code Llama 70B Instruct 43.6%
  17. 17 CodeLlama-70b 43.2%
  18. 18 DeepSeek-Coder-33b-base 42.8%
  19. 19 CodeLlama-70b-Python 42.3%
  20. 20 StarCoder 2 15B 39.9%
  21. 21 Code Llama 34B Instruct 38.6%
  22. 22 WizardCoder-Python-13B-V1.0 38.4%
  23. 23 DeepSeek-Coder-7b-base 38.3%
  24. 24 CodeLlama-34b 37.8%
  25. 25 CodeLlama-34b-Python 37%
  26. 26 WizardCoder-15B-V1.0 35.3%
  27. 27 CodeLlama-13b-Instruct 34.4%
  28. 28 CodeGemma-7B 33.8%
  29. 29 CodeLlama-13b 32.3%
  30. 30 CodeLlama-13b-Python 30.2%
  31. 31 CodeLlama-7b-Instruct 28.5%
  32. 32 StarCoder 2 7B 28.3%
  33. 33 CodeGemma-7B-it 27%
  34. 34 CodeShell-7B 26.9%
  35. 35 CodeLlama-7b 26.9%
  36. 36 OctoCoder-15B 25.3%
  37. 37 CodeLlama-7b-Python 25.1%
  38. 38 StarCoder-15B 24.6%
  39. 39 Falcon-180B 24.2%
  40. 40 StarCoderBase-15B 24%
  41. 41 StarCoder 2 3B 22.8%
  42. 42 CodeGeex2-6B 20.6%
  43. 43 StarCoderBase-7B 20.1%
  44. 44 OctoGeeX-7B 19.9%
  45. 45 WizardCoder-3B-V1.0 18.6%
  46. 46 CodeGen25-7B-multi 18.2%
  47. 47 CodeGemma-2B 17.5%
  48. 48 Refact-1.6B 17.2%
  49. 49 Stable-code-3b 16.2%
  50. 50 DeepSeek-Coder-1b-base 16%
  51. 51 StarCoderBase-3B 13.9%
  52. 52 WizardCoder-1B-V1.0 12.4%
  53. 53 Replit-2.7B 10.5%
  54. 54 CodeGen25-7B-mono 10.1%
  55. 55 StarCoderBase-1.1B 10%
  56. 56 CodeGen-16B-Multi 8.92%
  57. 57 StableCode-3B-alpha 7.88%
  58. 58 DeciCoder-1B 7.65%
  59. 59 Phi-1 6.92%
  60. 60 SantaCoder-1.1B 6.27%

Source: Official Big Code Models Leaderboard leaderboard

#EntryScoreSettingDateLinkSource
1 Qwen2.5-Coder-32B-Instruct 59.2% Nov 2024 huggingface.co Official leaderboard
2 Qwen2.5-Coder-32B 56.7% Sep 2024 huggingface.co Official leaderboard
3 OpenCodeInterpreter-DS-33B 33B 56.3% huggingface.co Official leaderboard
4 Nxcode-CQ-7B-orpo 7B 55.9% huggingface.co Official leaderboard
5 CodeQwen1.5-7B-Chat 7B 55.7% huggingface.co Official leaderboard
6 CodeFuse-DeepSeek-33b 33B 54.7% huggingface.co Official leaderboard
7 DeepSeek-Coder-33b-instruct 33B 52.3% huggingface.co Official leaderboard
8 Artigenz-Coder-DS-6.7B 6.7B 51.7% huggingface.co Official leaderboard
9 DeepSeek-Coder-7b-instruct 6.7B 50.6% huggingface.co Official leaderboard
10 OpenCodeInterpreter-DS-6.7B 6.7B 49.8% huggingface.co Official leaderboard
11 Phind-CodeLlama-34B-v2 34B 49.4% huggingface.co Official leaderboard
12 Phind-CodeLlama-34B-v1 34B 48.1% huggingface.co Official leaderboard
13 Phind-CodeLlama-34B-Python-v1 34B 46.6% huggingface.co Official leaderboard
14 CodeQwen1.5-7B 45.2% Apr 2024 huggingface.co Official leaderboard
15 WizardCoder-Python-34B-V1.0 34B 43.6% huggingface.co Official leaderboard
16 Code Llama 70B Instruct 43.6% Jan 2024 huggingface.co Official leaderboard
17 CodeLlama-70b 70B 43.2% huggingface.co Official leaderboard
18 DeepSeek-Coder-33b-base 33B 42.8% huggingface.co Official leaderboard
19 CodeLlama-70b-Python 70B 42.3% huggingface.co Official leaderboard
20 StarCoder 2 15B 39.9% Feb 2024 huggingface.co Official leaderboard
21 Code Llama 34B Instruct 38.6% Aug 2023 huggingface.co Official leaderboard
22 WizardCoder-Python-13B-V1.0 13B 38.4% huggingface.co Official leaderboard
23 DeepSeek-Coder-7b-base 6.7B 38.3% huggingface.co Official leaderboard
24 CodeLlama-34b 34B 37.8% huggingface.co Official leaderboard
25 CodeLlama-34b-Python 34B 37% huggingface.co Official leaderboard
26 WizardCoder-15B-V1.0 15B 35.3% huggingface.co Official leaderboard
27 CodeLlama-13b-Instruct 13B 34.4% huggingface.co Official leaderboard
28 CodeGemma-7B 7B 33.8% huggingface.co Official leaderboard
29 CodeLlama-13b 13B 32.3% huggingface.co Official leaderboard
30 CodeLlama-13b-Python 13B 30.2% huggingface.co Official leaderboard
31 CodeLlama-7b-Instruct 7B 28.5% huggingface.co Official leaderboard
32 StarCoder 2 7B 28.3% Feb 2024 huggingface.co Official leaderboard
33 CodeGemma-7B-it 7B 27% huggingface.co Official leaderboard
34 CodeShell-7B 7B 26.9% huggingface.co Official leaderboard
35 CodeLlama-7b 7B 26.9% huggingface.co Official leaderboard
36 OctoCoder-15B 15B 25.3% huggingface.co Official leaderboard
37 CodeLlama-7b-Python 7B 25.1% huggingface.co Official leaderboard
38 StarCoder-15B 15B 24.6% huggingface.co Official leaderboard
39 Falcon-180B 24.2% Sep 2023 huggingface.co Official leaderboard
40 StarCoderBase-15B 15B 24% huggingface.co Official leaderboard
41 StarCoder 2 3B 22.8% Feb 2024 huggingface.co Official leaderboard
42 CodeGeex2-6B 6B 20.6% huggingface.co Official leaderboard
43 StarCoderBase-7B 7B 20.1% huggingface.co Official leaderboard
44 OctoGeeX-7B 7B 19.9% huggingface.co Official leaderboard
45 WizardCoder-3B-V1.0 3B 18.6% huggingface.co Official leaderboard
46 CodeGen25-7B-multi 7B 18.2% huggingface.co Official leaderboard
47 CodeGemma-2B 2B 17.5% huggingface.co Official leaderboard
48 Refact-1.6B 1.6B 17.2% huggingface.co Official leaderboard
49 Stable-code-3b 3B 16.2% huggingface.co Official leaderboard
50 DeepSeek-Coder-1b-base 1B 16% huggingface.co Official leaderboard
51 StarCoderBase-3B 3B 13.9% huggingface.co Official leaderboard
52 WizardCoder-1B-V1.0 1.1B 12.4% huggingface.co Official leaderboard
53 Replit-2.7B 2.7B 10.5% huggingface.co Official leaderboard
54 CodeGen25-7B-mono 7B 10.1% huggingface.co Official leaderboard
55 StarCoderBase-1.1B 1.1B 10% huggingface.co Official leaderboard
56 CodeGen-16B-Multi 16B 8.92% huggingface.co Official leaderboard
57 StableCode-3B-alpha 3B 7.88% huggingface.co Official leaderboard
58 DeciCoder-1B 1B 7.65% huggingface.co Official leaderboard
59 Phi-1 1B 6.92% huggingface.co Official leaderboard
60 SantaCoder-1.1B 1.1B 6.27% huggingface.co Official leaderboard
#ModelScoreSettingDateLinkSource
1 Qwen2.5-Coder-32B-Instruct 59.2% Nov 2024 huggingface.co Official leaderboard
2 Qwen2.5-Coder-32B 56.7% Sep 2024 huggingface.co Official leaderboard
3 CodeQwen1.5-7B 45.2% Apr 2024 huggingface.co Official leaderboard
4 Code Llama 70B Instruct 43.6% Jan 2024 huggingface.co Official leaderboard
5 StarCoder 2 15B 39.9% Feb 2024 huggingface.co Official leaderboard
6 Code Llama 34B Instruct 38.6% Aug 2023 huggingface.co Official leaderboard
7 StarCoder 2 7B 28.3% Feb 2024 huggingface.co Official leaderboard
8 Falcon-180B 24.2% Sep 2023 huggingface.co Official leaderboard
9 StarCoder 2 3B 22.8% Feb 2024 huggingface.co Official leaderboard

Rankings by part

We read 12 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as humaneval-python, java and javascript. Qwen2.5-Coder-32B-Instruct leads 9 of them, and Qwen2.5-Coder-32B leads the other 3. Arrows show how far a model moved from its overall rank among the models we track.

humaneval-python

9 models we track have a score here.

  1. 1 Qwen2.5-Coder-32B-Instruct 83.2%
  2. 2 Code Llama 70B Instruct ↑2 75.6%
  3. 3 Qwen2.5-Coder-32B ↓1 57.1%
  4. 4 Code Llama 34B Instruct ↑2 50.8%
  5. 5 CodeQwen1.5-7B ↓2 50.8%
  6. 6 StarCoder 2 15B ↓1 44.1%
  7. 7 Falcon-180B ↑1 35.4%
  8. 8 StarCoder 2 7B ↓1 34.1%
  9. 9 StarCoder 2 3B 31.4%

5 models have held number one on the official leaderboard since its release in July 2023

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

Meta Other labs Qwen
Own leaderboard Not tested since Nov 2024

Source: Official Big Code Models Leaderboard 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 Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Qwen2.5-Coder-32B-Instruct CodeQwen1.5-7B Code Llama 34B Instruct 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Qwen2.5-Coder-32B-Instruct CodeQwen1.5-7B

Source: Official Big Code Models Leaderboard leaderboard

No lab model card reports it yet

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

More in 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%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • 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

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in November 2024, for Qwen2.5-Coder-32B-Instruct. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (60 entries).
  • Pass Models scored 9 models we track, plus 51 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in November 2024, for Qwen2.5-Coder-32B-Instruct.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Qwen2.5-Coder-32B (#188).
  • Pass Headroom The best is 59.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on huggingface.co: 9 rows for models we track and 51 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check