paper-with-me

Code Generation 벤치마크

Code Generation on PECC

8개 결과 · ⬇ CSV · JSON

Pass@3

3.1 9.242 15.38 21.53 27.67 2024-04 2026-09 Claude 3 Haiku — 27.67 (2024-04-29) GPT-3.5 Turbo — 23.75 (2024-04-29) codechat-bison — 11.39 (2024-04-29) chat-bison — 8.48 (2024-04-29) Mixtral-8x7B-Instruct — 8.35 (2024-04-29) Phi-3-mini-128k-instruct — 7.18 (2024-04-29) WizardLM-2-7B — 3.72 (2024-04-29) Llama-3-8B-Instruct — 3.1 (2024-04-29) Claude 3 Haiku — 27.67 (2024-04-29)
RankModel Pass@3 PaperCodeYear
1 Claude 3 Haiku 27.67 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
2 GPT-3.5 Turbo 23.75 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
3 codechat-bison 11.39 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
4 chat-bison 8.48 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
5 Mixtral-8x7B-Instruct 8.35 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
6 Phi-3-mini-128k-instruct 7.18 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
7 WizardLM-2-7B 3.72 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
8 Llama-3-8B-Instruct 3.1 PECC: Problem Extraction and Coding Challenges hallerpatrick/pecc 2024
1–8 / 8 페이지당 10 20 50 100