ICML-25-BenchName
Collection
All the resources for our BenchName benchmark! • 8 items • Updated
bleu float64 0.35 4.22 | chrf float64 11.9 36.6 | rouge1 float64 13.6 29.3 | rouge2 float64 2.03 7.66 | rougeL float64 9.75 20.5 | bertscore float64 0.82 0.86 | bertscore_normalized float64 -0.06 0.2 | model_name stringlengths 6 34 | model_availability stringclasses 8
values | model_url stringlengths 22 77 | urls stringclasses 2
values | context_size stringclasses 10
values | submitted_by stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
3.788207 | 35.762825 | 28.634027 | 6.599072 | 19.809862 | 0.862466 | 0.185103 | DeepSeek-V3 | DeepSeek-V3 license | https://github.com/deepseek-ai/DeepSeek-V3 | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
3.563112 | 34.832085 | 28.2486 | 6.51563 | 19.944066 | 0.86264 | 0.186131 | Llama-3.1-405B-Instruct | Llama-3.1 license | https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
3.634037 | 34.66037 | 27.624658 | 6.62563 | 19.267359 | 0.861101 | 0.177015 | Llama-3.1-70B-Instruct | Llama-3.1 license | https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
2.409405 | 31.02003 | 23.660841 | 4.767632 | 16.672768 | 0.853755 | 0.133489 | Llama-3.1-8B-Instruct | Llama-3.1 license | https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
2.107786 | 26.338474 | 21.04603 | 4.101617 | 15.147176 | 0.846142 | 0.088379 | Llama-3.2-3B-Instruct | Llama-3.2 license | https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
3.751216 | 33.542144 | 28.376241 | 6.414642 | 20.118623 | 0.86445 | 0.196858 | Llama-3.3-70B-Instruct | Llama-3.3 license | https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
0.529028 | 14.071807 | 14.660548 | 3.381467 | 10.264967 | 0.827523 | -0.021942 | QwQ-32B-Preview | Apache 2.0 license | https://qwenlm.github.io/blog/qwq-32b-preview/ | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 32768 | BenchName Team |
3.415088 | 33.739252 | 27.925645 | 6.038047 | 20.102536 | 0.861562 | 0.179743 | Qwen2.5-Coder-32B-Instruct | Apache 2.0 license | https://qwenlm.github.io/blog/qwen2.5-coder-family/ | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 131072 | BenchName Team |
1.957481 | 30.115595 | 21.012887 | 5.044542 | 14.382634 | 0.842955 | 0.0695 | Claude 3 Haiku | Proprietary | https://www.anthropic.com/news/claude-3-family | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 200000 | BenchName Team |
4.218904 | 36.590655 | 28.673284 | 7.656369 | 20.14355 | 0.858331 | 0.160599 | Claude 3 Opus | Proprietary | https://www.anthropic.com/news/claude-3-family | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 200000 | BenchName Team |
4.194774 | 34.846548 | 28.787354 | 6.134166 | 19.66573 | 0.862563 | 0.185673 | Claude 3.5 Sonnet | Proprietary | https://www.anthropic.com/news/3-5-models-and-computer-use | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 200000 | BenchName Team |
0.35484 | 11.862069 | 13.61453 | 2.63346 | 11.439077 | 0.845287 | 0.083314 | CodeT5 (fine-tuned for CMG) | Apache 2.0 license | https://huggingface.co/JetBrains-Research/cmg-codet5-without-history | [code](https://github.com/) | 512 | BenchName Team |
1.727289 | 23.09909 | 18.207055 | 3.641952 | 13.478552 | 0.843915 | 0.075187 | CodeLLaMA-13b-Instruct | Llama-2 license | https://huggingface.co/codellama/CodeLlama-13b-Instruct-hf | [code](https://github.com/) | 16000 | BenchName Team |
1.585776 | 24.631982 | 17.817213 | 3.684291 | 13.114157 | 0.843586 | 0.073235 | CodeLLaMA-34b-Instruct | Llama-2 license | https://huggingface.co/codellama/CodeLlama-34b-Instruct-hf | [code](https://github.com/) | 16000 | BenchName Team |
1.107509 | 26.637646 | 16.961141 | 2.807006 | 12.027552 | 0.834841 | 0.021418 | CodeLLaMA-7b-Instruct | Llama-2 license | https://huggingface.co/codellama/CodeLlama-7b-Instruct-hf | [code](https://github.com/) | 16000 | BenchName Team |
0.750153 | 22.449119 | 13.814553 | 2.029344 | 9.753143 | 0.821634 | -0.056834 | DeepSeek Coder 1.3b Instruct | DeepSeek license | https://huggingface.co/deepseek-ai/deepseek-coder-1.3b-instruct | [code](https://github.com/) | 16000 | BenchName Team |
1.742203 | 29.079802 | 21.010977 | 4.471145 | 14.458026 | 0.842528 | 0.066965 | DeepSeek Coder 33b Instruct | DeepSeek license | https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct | [code](https://github.com/) | 16000 | BenchName Team |
1.634172 | 28.567472 | 20.187917 | 3.60416 | 14.115834 | 0.842671 | 0.067816 | DeepSeek Coder 6.7b Instruct | DeepSeek license | https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-instruct | [code](https://github.com/) | 16000 | BenchName Team |
2.918087 | 34.636232 | 27.378328 | 5.865121 | 18.678649 | 0.858148 | 0.159514 | Gemini 1.5 Flash | Proprietary | https://ai.google.dev/gemini-api/docs/models/gemini#gemini-1.5-flash | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 1048576 | BenchName Team |
3.655983 | 34.868972 | 28.943994 | 6.362707 | 20.151534 | 0.85934 | 0.166578 | Gemini 1.5 Pro | Proprietary | https://ai.google.dev/gemini-api/docs/models/gemini#gemini-1.5-pro | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 2097152 | BenchName Team |
3.065907 | 34.806064 | 26.068198 | 5.547537 | 17.654481 | 0.854028 | 0.135104 | GPT-4o | Proprietary | https://openai.com/index/hello-gpt-4o/ | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
2.840511 | 34.123527 | 25.66269 | 5.158397 | 17.331376 | 0.857949 | 0.158338 | GPT-4o mini | Proprietary | https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/ | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
2.100729 | 26.663837 | 19.975573 | 4.226778 | 14.447362 | 0.845875 | 0.086797 | GPT-3.5 Turbo (0613) | Proprietary | https://platform.openai.com/docs/models/gpt-3-5 | [code](https://github.com/) | 16000 | BenchName Team |
1.885333 | 20.697809 | 18.42437 | 3.814669 | 14.087406 | 0.854107 | 0.135575 | GPT-3.5 Turbo (1106) | Proprietary | https://platform.openai.com/docs/models/gpt-3-5 | [code](https://github.com/) | 16000 | BenchName Team |
2.126868 | 32.624441 | 23.497446 | 5.21744 | 16.032817 | 0.852202 | 0.124288 | GPT-4 (0613) | Proprietary | https://openai.com/gpt-4 | [code](https://github.com/) | 8000 | BenchName Team |
2.803404 | 34.391045 | 26.621946 | 5.296136 | 17.717172 | 0.855898 | 0.146187 | GPT-4 Turbo (1106) | Proprietary | https://openai.com/blog/new-models-and-developer-products-announced-at-devday | [code](https://github.com/) | 128000 | BenchName Team |
1.894698 | 30.719033 | 23.648243 | 4.457864 | 16.261564 | 0.847482 | 0.096322 | Mistral-7b-Instruct-v0.2 | Apache 2.0 license | https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2 | [code](https://github.com/) | 16000 | BenchName Team |
2.188516 | 31.984321 | 23.609689 | 5.376493 | 16.328624 | 0.847577 | 0.096882 | Mixtral-8x7B-Instruct-v0.1 (8 bit) | Apache 2.0 license | https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1 | [code](https://github.com/) | 32000 | BenchName Team |
4.089698 | 34.327234 | 27.956597 | 6.712009 | 20.047533 | 0.860541 | 0.173693 | o1 mini | Proprietary | https://openai.com/o1/ | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
4.212471 | 36.377656 | 29.275511 | 7.660035 | 20.52308 | 0.863467 | 0.191031 | o1 preview | Proprietary | https://openai.com/o1/ | [code](https://anonymous.4open.science/r/icml-benchname-2025/) | 128000 | BenchName Team |
BenchName (raw results)
These are the raw results from the BenchName benchmark suite, as well as the corresponding model predictions.
Please use the subset dropdown menu to select the necessary data relating to our six benchmarks:
- 🤗 Library-based code generation
- 🤗 CI builds repair
- 🤗 Project-level code completion
- 🤗 Commit message generation
- 🤗 Bug localization
- 🤗 Module summarization
- Downloads last month
- 10