Codesota · Computer Vision · Optical Character Recognition · CodeSearchNetTasks/Computer Vision/Optical Character Recognition
Optical Character Recognition · benchmark dataset · 2020 · EN

CodeSearchNet.

Benchmark for code summarization (docstring generation) across 6 programming languages: Python, Java, JavaScript, PHP, Ruby, Go. Over 2M (code, docstring) pairs. Primary metric is BLEU-4.

Submit a result ↵
§ 01 · Leaderboard

Best published scores.

14 results indexed across 2 metrics. Shaded row marks current SOTA; ties broken by submission date.

Primary
accuracy · higher is better
All metrics
bleu-4, smoothed-bleu-4
bleu-4
7 rows
#ModelOrgSubmittedPaper / codebleu-4
01GPT-4oAPIOpenAIMar 2026arxiv25.30
02Qwen2.5-Coder 32BOpenAlibabaSep 2024Qwen2.5-Coder Technical Report · code23.40
03DeepSeek-Coder-V2-InstructOpenDeepSeekJun 2024DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source… · code22.80
04CodeT5+ 2BOpenSalesforceMay 2023CodeT5+: Open Code Large Language Models for Code Unders… · code21.36
05CodeT5+OpenSalesforceMay 2023CodeT5+: Open Code Large Language Models for Code Unders… · code20.01
06UniXcoderOpenMicrosoftMar 2022UniXcoder: Unified Cross-Modal Pre-Training for Code Rep… · code19.06
07CodeBERTOpenMicrosoftFeb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code17.65
smoothed-bleu-4
7 rows
#ModelOrgSubmittedPaper / codesmoothed-bleu-4
01CodeBERT (MLM+RTD)—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code15.99
02CodeBERT (MLM)—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code15.55
03pre-train w/ code only—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code15.15
04CodeBERT (RTD)—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code15.03
05RoBERTa—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code14.52
06Transformer—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code14.31
07seq2seq—Feb 2020CodeBERT: A Pre-Trained Model for Programming and Natura… · code13.36
Fig 2 · Rows sorted by score within each metric. Shaded row marks SOTA. Dates reflect model or paper release where available, otherwise the date Codesota accessed the source.
§ 04 · Literature

6 papers
tied to this benchmark.

Every paper below corresponds to at least one row in the leaderboard above. Click through for the arXiv preprint and, when available, the reference implementation.

§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result ↵Read submission guide
What a submission needs
  • 01A public checkpoint or API endpoint
  • 02A reproduction script with frozen commit + seed
  • 03Declared evaluation environment (Python, deps)
  • 04One row per metric declared by this dataset
  • 05A contact so we can follow up on discrepancies
CodeSearchNet — Optical Character Recognition | CodeSOTA