Skip to content

MLS-Bench Lite

A 30-task subset of MLS-Bench that evaluates whether AI systems can invent generalizable and scalable machine-learning methods.

4 models

Higher is better.

  1. 1Kimi K3Moonshot AI48.3
  2. 2Qwen3.8 MaxAlibaba41.0
  3. 3Step 5 PreviewStepFun40.5
  4. 4Kimi K2.7 CodeMoonshot AI35.1
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Agentic ML task evaluation
Tasks
30 machine-learning research tasks
Authors
MLS-Bench (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.