EvoEval
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%.
- Made by
- Chunqiu Steven Xia and colleagues
- Released
- March 2024
- Kind
- Function-level problems
- Level
- School
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- All public
- Measures
- Capability
- Language
- English
- 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
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:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 68.6 GPT-4 Turbo
- 68.6 GPT-4
- 67.4 Claude-3
- 60.5 DeepSeeker-33b-Inst
- 58.2 WizardCoder-33b-1.1
- 57.2 ChatGPT
- 57.0 Claude 3 Haiku
- 54.5 MagiCoder-s-ds-6.7b
- 54.1 Code-millenials-34b
- 54.0 Speechless-codellama-34b
- 53.2 DeepSeeker-6.7b-Inst
- 53.2 DeepSeeker-7b-Inst-v1.5
- 52.7 XwinCoder-34b
- 52.1 Phind-CodeLlamma-34b-v2
- 51.7 Claude 2
- 51.7
Code Llama 70B Instruct
- 48.8 Gemini
- 47.9 Openchat
- 47.5 Qwen-72b-1.5
- 46.8 MagiCoder-s-cl-7b
- 46.1 WizardCoder-34b
- 44.6 CodeLlamma-70b
- 42.1
DeepSeek Coder 33B
- 41.0 DeepSeeker-1.3b-Inst
- 38.7 Qwen-14b-1.5
- 37.7 DeepSeeker-7b-base-v1.5
- 37.7
DeepSeek Coder 6.7B
- 37.2 CodeLlamma-13b-Inst
- 36.6
Phi-2
- 36.3
Code Llama 34B Instruct
- 35.6 CodeLlamma-34b
- 35.5
Qwen-14B
- 32.8
StarCoder 2 15B
- 32.6 Mixtral-8x7b-Inst
- 32.5 CodeLlamma-13b
- 31.7 CodeLlamma-7b-Inst
- 30.0 PaLM-2
- 29.9 Qwen-7b-1.5
- 29.3 CodeLlamma-7b
- 27.9
StarCoder 2 7B
- 26.1 StarCoder
- 25.3
Gemma 7B
- 24.5
DeepSeek Coder 1.3B
- 24.2
Qwen-7B
- 23.4
StarCoder 2 3B
- 23.3 Stable-code-3b
- 22.7
Mistral 7B v0.1
- 20.2 Gemma-7b-Inst
- 20.1 Mistral-7b-Inst
- 17.4
Mistral-7B-Instruct-v0.2
- 15.2
Gemma 2B
- 1 GPT-4 Turbo Apr 2024 68.6%
- 2 GPT-4 Jun 2023 68.6%
- 3 Claude-3 67.4%
- 4 DeepSeeker-33b-Inst 60.5%
- 5 WizardCoder-33b-1.1 58.2%
- 6 ChatGPT 57.2%
- 7 Claude 3 Haiku 57%
- 8 MagiCoder-s-ds-6.7b 54.5%
- 9 Code-millenials-34b 54.1%
- 10 Speechless-codellama-34b 54%
- 11 DeepSeeker-6.7b-Inst 53.2%
- 12 DeepSeeker-7b-Inst-v1.5 53.2%
- 13 XwinCoder-34b 52.7%
- 14 Phind-CodeLlamma-34b-v2 52.1%
- 15 Claude 2 51.7%
- 16
Code Llama 70B Instruct 51.7%
- 17 Gemini 48.8%
- 18 Openchat 47.9%
- 19 Qwen-72b-1.5 47.5%
- 20 MagiCoder-s-cl-7b 46.8%
- 21 WizardCoder-34b 46.1%
- 22 CodeLlamma-70b 44.6%
- 23
DeepSeek Coder 33B 42.1%
- 24 DeepSeeker-1.3b-Inst 41%
- 25 Qwen-14b-1.5 38.7%
- 26 DeepSeeker-7b-base-v1.5 37.7%
- 27
DeepSeek Coder 6.7B 37.7%
- 28 CodeLlamma-13b-Inst 37.2%
- 29
Phi-2 36.6%
- 30
Code Llama 34B Instruct 36.3%
- 31 CodeLlamma-34b 35.6%
- 32
Qwen-14B 35.5%
- 33
StarCoder 2 15B 32.8%
- 34 Mixtral-8x7b-Inst 32.6%
- 35 CodeLlamma-13b 32.5%
- 36 CodeLlamma-7b-Inst 31.7%
- 37 PaLM-2 30%
- 38 Qwen-7b-1.5 29.9%
- 39 CodeLlamma-7b 29.3%
- 40
StarCoder 2 7B 27.9%
- 41 StarCoder 26.1%
- 42
Gemma 7B 25.3%
- 43
DeepSeek Coder 1.3B 24.5%
- 44
Qwen-7B 24.2%
- 45
StarCoder 2 3B 23.4%
- 46 Stable-code-3b 23.3%
- 47
Mistral 7B v0.1 22.7%
- 48 Gemma-7b-Inst 20.2%
- 49 Mistral-7b-Inst 20.1%
- 50
Mistral-7B-Instruct-v0.2 17.4%
- 51
Gemma 2B 15.2%
Source: Official EvoEval leaderboard
- 68.6 GPT-4 Turbo
- 68.6 GPT-4
- 57.0 Claude 3 Haiku
- 51.7 Claude 2
- 51.7
Code Llama 70B Instruct
- 42.1
DeepSeek Coder 33B
- 37.7
DeepSeek Coder 6.7B
- 36.6
Phi-2
- 36.3
Code Llama 34B Instruct
- 35.5
Qwen-14B
- 32.8
StarCoder 2 15B
- 27.9
StarCoder 2 7B
- 25.3
Gemma 7B
- 24.5
DeepSeek Coder 1.3B
- 24.2
Qwen-7B
- 23.4
StarCoder 2 3B
- 22.7
Mistral 7B v0.1
- 17.4
Mistral-7B-Instruct-v0.2
- 15.2
Gemma 2B
- 1 GPT-4 Turbo Apr 2024 68.6%
- 2 GPT-4 Jun 2023 68.6%
- 3 Claude 3 Haiku 57%
- 4 Claude 2 51.7%
- 5
Code Llama 70B Instruct 51.7%
- 6
DeepSeek Coder 33B 42.1%
- 7
DeepSeek Coder 6.7B 37.7%
- 8
Phi-2 36.6%
- 9
Code Llama 34B Instruct 36.3%
- 10
Qwen-14B 35.5%
- 11
StarCoder 2 15B 32.8%
- 12
StarCoder 2 7B 27.9%
- 13
Gemma 7B 25.3%
- 14
DeepSeek Coder 1.3B 24.5%
- 15
Qwen-7B 24.2%
- 16
StarCoder 2 3B 23.4%
- 17
Mistral 7B v0.1 22.7%
- 18
Mistral-7B-Instruct-v0.2 17.4%
- 19
Gemma 2B 15.2%
Source: Official EvoEval leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 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 | | 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 | | 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 | | 37.7% | Nov 2023 | Official leaderboard | |
| 28 | CodeLlamma-13b-Inst | 37.2% | Official leaderboard | ||
| 29 | | 36.6% | Dec 2023 | Official leaderboard | |
| 30 | | 36.3% | Aug 2023 | Official leaderboard | |
| 31 | CodeLlamma-34b | 35.6% | Official leaderboard | ||
| 32 | | 35.5% | Sep 2023 | Official leaderboard | |
| 33 | | 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 | | 27.9% | Feb 2024 | Official leaderboard | |
| 41 | StarCoder | 26.1% | Official leaderboard | ||
| 42 | | 25.3% | Feb 2024 | Official leaderboard | |
| 43 | | 24.5% | Nov 2023 | Official leaderboard | |
| 44 | | 24.2% | Sep 2023 | Official leaderboard | |
| 45 | | 23.4% | Feb 2024 | Official leaderboard | |
| 46 | Stable-code-3b | 23.3% | Official leaderboard | ||
| 47 | | 22.7% | Sep 2023 | Official leaderboard | |
| 48 | Gemma-7b-Inst | 20.2% | Official leaderboard | ||
| 49 | Mistral-7b-Inst | 20.1% | Official leaderboard | ||
| 50 | | 17.4% | Dec 2023 | Official leaderboard | |
| 51 | | 15.2% | Feb 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 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 | | 51.7% | Jan 2024 | Official leaderboard | |
| 6 | | 42.1% | Nov 2023 | Official leaderboard | |
| 7 | | 37.7% | Nov 2023 | Official leaderboard | |
| 8 | | 36.6% | Dec 2023 | Official leaderboard | |
| 9 | | 36.3% | Aug 2023 | Official leaderboard | |
| 10 | | 35.5% | Sep 2023 | Official leaderboard | |
| 11 | | 32.8% | Feb 2024 | Official leaderboard | |
| 12 | | 27.9% | Feb 2024 | Official leaderboard | |
| 13 | | 25.3% | Feb 2024 | Official leaderboard | |
| 14 | | 24.5% | Nov 2023 | Official leaderboard | |
| 15 | | 24.2% | Sep 2023 | Official leaderboard | |
| 16 | | 23.4% | Feb 2024 | Official leaderboard | |
| 17 | | 22.7% | Sep 2023 | Official leaderboard | |
| 18 | | 17.4% | Dec 2023 | Official leaderboard | |
| 19 | | 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 GPT-4 Turbo Apr 2024 83.5%
- 2 GPT-4 Jun 2023 82.3%
- 3 Claude 3 Haiku 74.4%
- 4 Claude 2 66.5%
- 5
Code Llama 70B Instruct 66.5%
- 6
Code Llama 34B Instruct ↑3 51.8%
- 7
DeepSeek Coder 33B ↓1 50.6%
- 8
Phi-2 50%
- 9
Qwen-14B ↑1 46.3%
- 10
DeepSeek Coder 6.7B ↓3 45.1%
Difficult
19 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 52%
- 2 GPT-4 Turbo Apr 2024 ↓1 50%
- 3 Claude 3 Haiku 40%
- 4
Code Llama 70B Instruct ↑1 31%
- 5 Claude 2 ↓1 29%
- 6
DeepSeek Coder 33B 26%
- 7
Qwen-14B ↑3 23%
- 8
Code Llama 34B Instruct ↑1 22%
- 9
DeepSeek Coder 6.7B ↓2 21%
- 10
Phi-2 ↓2 18%
Creative
19 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 66%
- 2 GPT-4 Turbo Apr 2024 ↓1 61%
- 3 Claude 3 Haiku 47%
- 4 Claude 2 42%
- 5
Code Llama 70B Instruct 41%
- 6
Code Llama 34B Instruct ↑3 27%
- 7
DeepSeek Coder 6.7B 24%
- 8
DeepSeek Coder 33B ↓2 23%
- 9
Phi-2 ↓1 23%
- 10
Qwen-14B 20%
Subtle
19 models we track have a score here.
- 1 GPT-4 Turbo Apr 2024 82%
- 2 GPT-4 Jun 2023 76%
- 3
Code Llama 70B Instruct ↑2 65%
- 4 Claude 3 Haiku ↓1 65%
- 5 Claude 2 ↓1 64%
- 6
Phi-2 ↑2 49%
- 7
DeepSeek Coder 33B ↓1 47%
- 8
DeepSeek Coder 6.7B ↓1 47%
- 9
Qwen-14B ↑1 45%
- 10
Code Llama 34B Instruct ↓1 43%
Combine
19 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 53%
- 2 GPT-4 Turbo Apr 2024 ↓1 45%
- 3 Claude 3 Haiku 25%
- 4 Claude 2 19%
- 5
Code Llama 70B Instruct 18%
- 6
Phi-2 ↑2 14%
- 7
Qwen-14B ↑3 13%
- 8
DeepSeek Coder 33B ↓2 11%
- 9
Code Llama 34B Instruct 9.00%
- 10
DeepSeek Coder 6.7B ↓3 5.00%
Tool Use
19 models we track have a score here.
- 1 GPT-4 Turbo Apr 2024 69%
- 2 GPT-4 Jun 2023 68%
- 3
Code Llama 70B Instruct ↑2 65%
- 4
DeepSeek Coder 33B ↑2 63%
- 5 Claude 3 Haiku ↓2 61%
- 6 Claude 2 ↓2 57%
- 7
DeepSeek Coder 6.7B 55%
- 8
StarCoder 2 15B ↑3 48%
- 9
Code Llama 34B Instruct 47%
- 10
StarCoder 2 7B ↑2 46%
Verbose
19 models we track have a score here.
- 1 GPT-4 Turbo Apr 2024 79.3%
- 2 Claude 3 Haiku ↑1 71.3%
- 3 GPT-4 Jun 2023 ↓1 70.1%
- 4 Claude 2 66.5%
- 5
Code Llama 70B Instruct 65.9%
- 6
DeepSeek Coder 33B 61%
- 7
DeepSeek Coder 6.7B 57.3%
- 8
Phi-2 51.2%
- 9
Qwen-14B ↑1 48.8%
- 10
StarCoder 2 15B ↑1 46.3%
Concise
19 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 81.1%
- 2 GPT-4 Turbo Apr 2024 ↓1 79.3%
- 3 Claude 3 Haiku 72%
- 4 Claude 2 69.5%
- 5
Code Llama 70B Instruct 61%
- 6
DeepSeek Coder 33B 54.9%
- 7
Phi-2 ↑1 50.6%
- 8
Code Llama 34B Instruct ↑1 50%
- 9
DeepSeek Coder 6.7B ↓2 47%
- 10
Qwen-14B 43.3%
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.
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.
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
- 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.