A smaller model. A better fit.
Model discovery becomes difficult when a single leaderboard stands in for every task. Function calling, extraction, document understanding and classification ask different things of a model.
The approach
slm.expert organizes small open-weight models by use case. Model pages bring together published benchmark results, licenses and platform information, with links back to the evidence. It gives builders a place to narrow their options before testing in their own workflow.
- 1Choose a use case
- 2Compare evidence
- 3Test your shortlist
- Task-oriented categories cover areas such as tool use, structured extraction and document AI.
- Published scores are presented with their sources so comparisons can be checked.
- License information is part of the model decision, alongside capability and size.
Explore & contribute
Explore the directory and bring corrections, missing models or useful evidence into the conversation. This is a resource about open-weight models; each model’s own license still applies.