---
title: Fabryka AI track leaderboard
emoji: 🏁
colorFrom: gray
colorTo: blue
sdk: static
pinned: false
license: apache-2.0
short_description: Fabryka AI leaderboard of small language models
tags:
  - leaderboard
models:
  - Maggio33/GoLLeM-110M-PL
  - Maggio33/GoLLeM-110M-PL-v2
  - Maggio33/GoLLeM-110M-PL-v3
  - SlayerLab/GoLLeM-110M-PL-v2
  - SlayerLab/GoLLeM-110M-PL-v3
  - Maggio33/gollem-v4-250m-pl
  - SlayerLab/gollem-v4-250m-pl
  - SlayerLab/GoLLeM-45M-PL
  - SlayerLab/pollock-mini-lm-125m
  - SlayerLab/Slayer149-Balanced-ARC-E-ft-45500
  - SlayerLab/gollem-v5-ckpts
  - SlayerLab/GoLLeM-v6-250M
  - Maggio33/gollem-v6-250m
  - Maggio33/Slayer139-1.01
  - SlayerLab/Slayer139-1.01
---

# Fabryka AI track leaderboard

The models listed in this Space's metadata have a row on the Fabryka AI model leaderboard:
<https://track.fabryka.ai/models>

The leaderboard ranks small language models with the Glint Tiny-ML protocol (BLiMP, ARC-Easy and
WikiText-2 byte perplexity, combined into the "eff" score; the size bonus in eff is scaled for 1K–150M
parameters). Each row states whether the numbers were measured by Fabryka or reported by the model's authors.