BREAKING
Grug-12B Cuts Reasoning Tokens ~70%
0
%
fewer tokens
0
%
quality gap
0
B
parameters
Tokens Per Response
Gemma-4
228.53
Grug-12B
68.94
Built Cheaply on Open Weights
1
Base Gemma-4-12B
↓
2
QLoRA on 5,740 rows
↓
3
35 min on one A100
↓
4
Merged safetensors
Speed Gains vs Reliability Risk
Strengths
~40% faster
●
34 vs 24 tokens/sec
●
Good for agents and scraping
●
1,134+ downloads
Limits
experimental
●
Unstable on multi-hop
●
~47% pass@1 coding
●
Format not always suppressed
A Promising Proof of Concept
AI NEWS BLITZ
An open-source fine-tune of Gemma-4 slashes thinking tokens while nearly matching base quality.