MLE-bench Lite
MLE-bench Lite - machine-learning-engineering agent benchmark (iteratively train, evaluate and improve ML systems); Lite is the reduced subset of OpenAI's MLE-bench. Scores are pass rates on MLE-bench Lite tasks; official repo: https://github.com/openai/mle-bench (paper: https://arxiv.org/abs/2410.07095).
Category
Agentic Coding
Score type
percent
Scale
100%
higher is better
Models tested
3
What is MLE-bench Lite?
MLE-bench Lite is a benchmark for how well a model works as an autonomous coding agent. The official description says: "MLE-bench Lite - machine-learning-engineering agent benchmark (iteratively train, evaluate and improve ML systems); Lite is the reduced subset of OpenAI's MLE-bench". Below we describe in simple words what benchmarks of this type generally examine.
What does it test?
MLE-bench Lite measures how well a model works as an autonomous coding agent. In detail, it checks:
- fixing real, open bugs from real software repositories
- planning multi-step changes, running commands and editing files by itself
- passing the project's test suite after its changes
- handling whole repositories, not just small isolated code snippets
How to read the score
The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.
Official methodology note: MLE-bench Lite - machine-learning-engineering agent benchmark (iteratively train, evaluate and improve ML systems); Lite is the reduced subset of OpenAI's MLE-bench. Scores are pass rates on MLE-bench Lite tasks; official repo: https://github.com/openai/mle-bench (paper: https://arxiv.org/abs/2410.07095).
Frequently asked questions
What is the MLE-bench Lite benchmark?
MLE-bench Lite is a benchmark for how well a model works as an autonomous coding agent. The official description says: 'MLE-bench Lite - machine-learning-engineering agent benchmark (iteratively train, evaluate and improve ML systems); Lite is the reduced subset of OpenAI's MLE-bench'. Below we describe in simple words what benchmarks of this type generally examine.
What does MLE-bench Lite test?
MLE-bench Lite mainly measures how well a model works as an autonomous coding agent. Concretely, it checks: fixing real, open bugs from real software repositories; planning multi-step changes, running commands and editing files by itself; passing the project's test suite after its changes.
How is the MLE-bench Lite score calculated?
The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better. The declared maximum score of this benchmark is 100.0.
Where do I get MLE-bench Lite data and scores?
The official source is https://github.com/openai/mle-bench. The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/mle-bench-lite/scores) or pull it directly from the Open LLM Radar API with any HTTP client.
Benchmark data & downloads
Where to get, inspect and reuse the data behind this benchmark.
Official benchmark page / paper
The benchmark maintainer's website, paper or leaderboard
Benchmark metadata (JSON)
This page's benchmark record from the Open LLM Radar API
All model scores (JSON download)
Full score table as JSON; paginate with limit/offset params
Open LLM Radar API docs
Swagger UI to explore and reuse all radar API endpoints
Model scores 3 models tested
| # | Model | Score | Date | Source | Verified |
|---|---|---|---|---|---|
| 1 |
Sage-S2.1-Flash
Shanghai AI Laboratory
|
90.3%
|
— | Report | Verified |
| 2 |
DeepSeek-V4.1-Flash-1017
DeepSeek
|
89.2%
|
— | Report | Unverified |
| 3 |
GLM-5.3-Flash
Zhipu AI
|
84.3%
|
— | Report | Unverified |