Skip to content

Complexity Hurts Small Models

On qwen2.5-coder:1.5b, increasing prompt complexity from minimal to maximal dropped pass rate from 0.78 to 0.28.

Complexity by model

Key Numbers

Model minimal role + constraints examples maximal
qwen2.5-coder:1.5b 0.78 0.94 0.83 0.28
gemma3:1b 0.94 0.94 0.83 0.83
phi4-mini 0.89 0.94 0.72 0.94
gemma3:4b 0.94 0.94 0.94 0.94

The strongest negative effect appears on the smallest coder model. Larger tested models were mostly flat.

Reproduce

The public repo ships the derived summary in data/public/findings.json. The contribution harness currently supports E9 first; E8 reproduction commands will be promoted after the harness migration is complete.

Limitations

Complexity is represented as four discrete prompt templates, not a continuous scale. The result should be read as a capacity warning for small models, not a universal rule to make every prompt shorter.