CodeXGLUE
A broad code-intelligence suite spanning understanding, retrieval, completion, translation, repair, generation, and summarization.
The official challenge and repository remain available, but the heterogeneous task-specific leaderboard is no longer a current frontier-code comparison and has no maintained common aggregate.
Performance Timeline
Longitudinal progression of model scores against human baselines.Performance & Historical Trajectory
Empirical score progression across model release dates and evaluation rounds.
| Model | Score | Date | Source Type | Provenance |
|---|---|---|---|---|
| Seq2Seq (CodeSearchNet summarization, six-language overall BLEU) | 14.32% | 2021-02-09 | independent | Source ↗ |
| Transformer (CodeSearchNet summarization, six-language overall BLEU) | 15.56% | 2021-02-09 | independent | Source ↗ |
| RoBERTa encoder (CodeSearchNet summarization, six-language overall BLEU) | 16.57% | 2021-02-09 | independent | Source ↗ |
| CodeBERT encoder (CodeSearchNet summarization, six-language overall BLEU) | 17.83% | 2021-02-09 | vendor-reported | Source ↗ |
Human Baseline & Difficulty Horizon
Calibrated human reference points, specialist benchmarks, and ceiling thresholds.Developer reference baseline across code comprehension, translation, and synthesis tasks.
Metric & Scoring Methodology
Verification protocols, aggregation formulas, and specialized metric variants.overall smoothed BLEU on six-language CodeSearchNet code summarizationDataset & Compute Cost
Evaluation volume, public availability, API pricing, and local hardware requirements.$5 – $20 USD for full benchmark evaluation run on frontier APIs.
1x NVIDIA RTX 4090 (24GB) or A100 (40GB/80GB) via vLLM / SGLang
How to Run & Reproduce
Standardized evaluation protocols, CLI commands, and reproducible runner templates.lm_eval --model hf --model_args pretrained=<model_path> --tasks codexglue --batch_size autoopencompass --datasets codexglue --models <model_config># Standard API Evaluation Loop
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gpt-4o",
messages=[{"role": "user", "content": prompt}],
temperature=0.0,
)When publishing results for CodeXGLUE, always report the exact prompt template, few-shot exemplar ordering, sampling temperature (temperature=0), maximum reasoning budget tokens, and the precise timestamped model snapshot ID.
Contamination & Memorization Analysis
Audit of pretraining exposure risks, memorization vectors, and refresh policies.Static fixed snapshot
Public on web / HuggingFace