Benchmark Almanac

EvoEval

GPT-4 Turbo Apr 2024 68.6% leads since April 2024
Not in model cards

Models solve Python problems that vary the constraints, interfaces, or logic of HumanEval tasks. Harbor ports the 100 problems in the Difficult subset.

The leaderboard listed 51 models on November 19, 2026. GPT-4 Turbo led overall at 68.63%, and GPT-4 led the Difficult subset at 52%.

Released
March 2024
Kind
Function-level problems
Level
School
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Code
Interaction
Single turn
Items from
Another benchmark
Refresh
One fixed set
Run by
Anyone
Based on
HumanEval
Strongest model tested
GPT-4 Turbo Apr 2024 #172 on the capability index
Last model added
April 2024
Scores refreshed
November 20th, 2026
51
Entries on its own leaderboard, 19 models we track

GPT-4 Turbo leads with 68.6% among models released by April 2024 GPT-4 Turbo leads with 68.6% among models released by April 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 April 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-4 Turbo Apr 2024 68.6%
  2. 2 GPT-4 Jun 2023 68.6%
  3. 3 Claude-3 67.4%
  4. 4 DeepSeeker-33b-Inst 60.5%
  5. 5 WizardCoder-33b-1.1 58.2%
  6. 6 ChatGPT 57.2%
  7. 7 Claude 3 Haiku 57%
  8. 8 MagiCoder-s-ds-6.7b 54.5%
  9. 9 Code-millenials-34b 54.1%
  10. 10 Speechless-codellama-34b 54%
  11. 11 DeepSeeker-6.7b-Inst 53.2%
  12. 12 DeepSeeker-7b-Inst-v1.5 53.2%
  13. 13 XwinCoder-34b 52.7%
  14. 14 Phind-CodeLlamma-34b-v2 52.1%
  15. 15 Claude 2 51.7%
  16. 16 Code Llama 70B Instruct 51.7%
  17. 17 Gemini 48.8%
  18. 18 Openchat 47.9%
  19. 19 Qwen-72b-1.5 47.5%
  20. 20 MagiCoder-s-cl-7b 46.8%
  21. 21 WizardCoder-34b 46.1%
  22. 22 CodeLlamma-70b 44.6%
  23. 23 DeepSeek Coder 33B 42.1%
  24. 24 DeepSeeker-1.3b-Inst 41%
  25. 25 Qwen-14b-1.5 38.7%
  26. 26 DeepSeeker-7b-base-v1.5 37.7%
  27. 27 DeepSeek Coder 6.7B 37.7%
  28. 28 CodeLlamma-13b-Inst 37.2%
  29. 29 Phi-2 36.6%
  30. 30 Code Llama 34B Instruct 36.3%
  31. 31 CodeLlamma-34b 35.6%
  32. 32 Qwen-14B 35.5%
  33. 33 StarCoder 2 15B 32.8%
  34. 34 Mixtral-8x7b-Inst 32.6%
  35. 35 CodeLlamma-13b 32.5%
  36. 36 CodeLlamma-7b-Inst 31.7%
  37. 37 PaLM-2 30%
  38. 38 Qwen-7b-1.5 29.9%
  39. 39 CodeLlamma-7b 29.3%
  40. 40 StarCoder 2 7B 27.9%
  41. 41 StarCoder 26.1%
  42. 42 Gemma 7B 25.3%
  43. 43 DeepSeek Coder 1.3B 24.5%
  44. 44 Qwen-7B 24.2%
  45. 45 StarCoder 2 3B 23.4%
  46. 46 Stable-code-3b 23.3%
  47. 47 Mistral 7B v0.1 22.7%
  48. 48 Gemma-7b-Inst 20.2%
  49. 49 Mistral-7b-Inst 20.1%
  50. 50 Mistral-7B-Instruct-v0.2 17.4%
  51. 51 Gemma 2B 15.2%

Source: Official EvoEval leaderboard

#EntryScoreSettingDateSource
1 GPT-4 Turbo Apr 2024 68.6% Apr 2024 Official leaderboard
2 GPT-4 Jun 2023 68.6% Jun 2023 Official leaderboard
3 Claude-3 67.4% Official leaderboard
4 DeepSeeker-33b-Inst 60.5% Official leaderboard
5 WizardCoder-33b-1.1 58.2% Official leaderboard
6 ChatGPT 57.2% Official leaderboard
7 Claude 3 Haiku 57% Mar 2024 Official leaderboard
8 MagiCoder-s-ds-6.7b 54.5% Official leaderboard
9 Code-millenials-34b 54.1% Official leaderboard
10 Speechless-codellama-34b 54% Official leaderboard
11 DeepSeeker-6.7b-Inst 53.2% Official leaderboard
12 DeepSeeker-7b-Inst-v1.5 53.2% Official leaderboard
13 XwinCoder-34b 52.7% Official leaderboard
14 Phind-CodeLlamma-34b-v2 52.1% Official leaderboard
15 Claude 2 51.7% Jul 2023 Official leaderboard
16 Code Llama 70B Instruct 51.7% Jan 2024 Official leaderboard
17 Gemini 48.8% Official leaderboard
18 Openchat 47.9% Official leaderboard
19 Qwen-72b-1.5 47.5% Official leaderboard
20 MagiCoder-s-cl-7b 46.8% Official leaderboard
21 WizardCoder-34b 46.1% Official leaderboard
22 CodeLlamma-70b 44.6% Official leaderboard
23 DeepSeek Coder 33B 42.1% Nov 2023 Official leaderboard
24 DeepSeeker-1.3b-Inst 41% Official leaderboard
25 Qwen-14b-1.5 38.7% Official leaderboard
26 DeepSeeker-7b-base-v1.5 37.7% Official leaderboard
27 DeepSeek Coder 6.7B 37.7% Nov 2023 Official leaderboard
28 CodeLlamma-13b-Inst 37.2% Official leaderboard
29 Phi-2 36.6% Dec 2023 Official leaderboard
30 Code Llama 34B Instruct 36.3% Aug 2023 Official leaderboard
31 CodeLlamma-34b 35.6% Official leaderboard
32 Qwen-14B 35.5% Sep 2023 Official leaderboard
33 StarCoder 2 15B 32.8% Feb 2024 Official leaderboard
34 Mixtral-8x7b-Inst 32.6% Official leaderboard
35 CodeLlamma-13b 32.5% Official leaderboard
36 CodeLlamma-7b-Inst 31.7% Official leaderboard
37 PaLM-2 30% Official leaderboard
38 Qwen-7b-1.5 29.9% Official leaderboard
39 CodeLlamma-7b 29.3% Official leaderboard
40 StarCoder 2 7B 27.9% Feb 2024 Official leaderboard
41 StarCoder 26.1% Official leaderboard
42 Gemma 7B 25.3% Feb 2024 Official leaderboard
43 DeepSeek Coder 1.3B 24.5% Nov 2023 Official leaderboard
44 Qwen-7B 24.2% Sep 2023 Official leaderboard
45 StarCoder 2 3B 23.4% Feb 2024 Official leaderboard
46 Stable-code-3b 23.3% Official leaderboard
47 Mistral 7B v0.1 22.7% Sep 2023 Official leaderboard
48 Gemma-7b-Inst 20.2% Official leaderboard
49 Mistral-7b-Inst 20.1% Official leaderboard
50 Mistral-7B-Instruct-v0.2 17.4% Dec 2023 Official leaderboard
51 Gemma 2B 15.2% Feb 2024 Official leaderboard
#ModelScoreSettingDateSource
1 GPT-4 Turbo Apr 2024 68.6% Apr 2024 Official leaderboard
2 GPT-4 Jun 2023 68.6% Jun 2023 Official leaderboard
3 Claude 3 Haiku 57% Mar 2024 Official leaderboard
4 Claude 2 51.7% Jul 2023 Official leaderboard
5 Code Llama 70B Instruct 51.7% Jan 2024 Official leaderboard
6 DeepSeek Coder 33B 42.1% Nov 2023 Official leaderboard
7 DeepSeek Coder 6.7B 37.7% Nov 2023 Official leaderboard
8 Phi-2 36.6% Dec 2023 Official leaderboard
9 Code Llama 34B Instruct 36.3% Aug 2023 Official leaderboard
10 Qwen-14B 35.5% Sep 2023 Official leaderboard
11 StarCoder 2 15B 32.8% Feb 2024 Official leaderboard
12 StarCoder 2 7B 27.9% Feb 2024 Official leaderboard
13 Gemma 7B 25.3% Feb 2024 Official leaderboard
14 DeepSeek Coder 1.3B 24.5% Nov 2023 Official leaderboard
15 Qwen-7B 24.2% Sep 2023 Official leaderboard
16 StarCoder 2 3B 23.4% Feb 2024 Official leaderboard
17 Mistral 7B v0.1 22.7% Sep 2023 Official leaderboard
18 Mistral-7B-Instruct-v0.2 17.4% Dec 2023 Official leaderboard
19 Gemma 2B 15.2% Feb 2024 Official leaderboard

Rankings by part

We read 8 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as HumanEval, Difficult and Creative. GPT-4 Turbo leads 4 of them, and GPT-4 leads the other 4. Arrows show how far a model moved from its overall rank among the models we track.

HumanEval

19 models we track have a score here.

  1. 1 GPT-4 Turbo Apr 2024 83.5%
  2. 2 GPT-4 Jun 2023 82.3%
  3. 3 Claude 3 Haiku 74.4%
  4. 4 Claude 2 66.5%
  5. 5 Code Llama 70B Instruct 66.5%
  6. 6 Code Llama 34B Instruct ↑3 51.8%
  7. 7 DeepSeek Coder 33B ↓1 50.6%
  8. 8 Phi-2 50%
  9. 9 Qwen-14B ↑1 46.3%
  10. 10 DeepSeek Coder 6.7B ↓3 45.1%

2 models have held number one on the official leaderboard since its release in March 2024

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

OpenAI
Own leaderboard Not tested since Apr 2024
GPT-4 Turbo 68.6%

Source: Official EvoEval 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% 2023 2024 Near the ceiling (85%) GPT-4 Turbo GPT-4 0% 20% 40% 60% 80% 100% 2023 2024 Near the ceiling (85%) GPT-4 Turbo GPT-4

Source: Official EvoEval 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 April 2024, for GPT-4 Turbo. 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 (51 entries).
  • Pass Models scored 19 models we track, plus 32 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in April 2024, for GPT-4 Turbo.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-4 Turbo (#172).
  • Pass Headroom The best is 68.6%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on evo-eval.github.io: 19 rows for models we track and 32 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