---
title: Open SLM Leaderboard
emoji: 🏆
colorFrom: red
colorTo: yellow
sdk: static
pinned: true
license: other
short_description: Open Small Language Model Leaderboard
---

## Updating leaderboard history

Models in `index.html` store an `addedDate` (`YYYY-MM-DD`) for their first
leaderboard appearance. This is metadata only and is not shown in the table.
The history charts cover the <150M, <100M, <50M, <10M and <3M categories.

Charts calculate directly in the browser from currently uncommented `MODELS`,
their addition dates, current Intelligence Index scores, parameters and organizations.
Each model enters on its addition date; the highest current II among models
available on that day determines the leader. Totals include today in Sydney.
Score and model edits update the charts automatically on page reload.
Exact score ties follow the order of `MODELS`.

Hugging Face, Meta, OpenAI, Google and Microsoft are excluded from these charts.
Size categories overlap, so a model or lab can earn multiple class-days on a
single calendar day. Commented-out models are excluded throughout the timeline.

When adding a model, set its `addedDate`. To backfill missing dates from local
Git history, run:

```powershell
python scripts/update_history.py
```

The script preserves existing dates and traces unambiguous model renames and
owner moves with unchanged parameters and benchmark scores. For models absent
from Git history, it uses today's Sydney date, or an explicit fallback:

```powershell
python scripts/update_history.py --added-date 2026-10-14
```

Git history is used only to backfill missing dates, never to calculate chart rankings.