Running
1
MINT
đż
Quantize LLMs to fit a memory budget
AI & ML interests
Model Quantization
Recent Activity
View all activity
Organization Card
baa.ai
Smaller. Smarter. Sovereign.
Making frontier models run anywhere
We build open tools for efficient AI deployment. Our research focuses on quantization methods that preserve model quality while dramatically reducing hardware requirements â bringing 400B+ parameter models to a single machine.
baa.aiâ¡âMINT Paperâ¡âGitHub
spaces 5
Running
1
SWAN: Data-Free Mixed-Precision Quantization
đŚ˘
Quantize LLMs without data using perâtensor mixed precision
Running
2
SAKD: SWAN-Guided Knowledge Distillation
đ
Generate quantizationâready student models via guided distillation
Running
1
Sensitivity-Aware Training (SAT)
đ
Train LLMs to be quantizationâready with sensitivityâaware methods
models 34
baa-ai/Llama-4-Maverick-17B-128E-Instruct-MINT-407GB-GGUF
Text Generation ⢠401B ⢠Updated ⢠11
baa-ai/MiniMax-M2.5-MINT-179GB-GGUF
Text Generation ⢠229B ⢠Updated ⢠17
baa-ai/GLM-5-SWAN-5bit-MLX
744B ⢠Updated ⢠480
baa-ai/Llama-3.1-70B-Instruct-SWAN-5bit-MLX
71B ⢠Updated ⢠232
baa-ai/Llama-3.3-70B-Instruct-SWAN-5bit-MLX
71B ⢠Updated ⢠274
baa-ai/Llama-4-Scout-17B-16E-Instruct-SWAN-4bit-MLX
18B ⢠Updated ⢠213
baa-ai/Llama-4-Maverick-17B-128E-Instruct-SWAN-4bit-MLX
51B ⢠Updated ⢠200
baa-ai/MiniMax-M2.5-SWAN-4bit-MLX
229B ⢠Updated ⢠735
baa-ai/GLM-4.7-Flash-SWAN-4bit-MLX
30B ⢠Updated ⢠526 ⢠1
baa-ai/Qwen3-8B-SWAN-6bit-MLX
8B ⢠Updated ⢠80 ⢠1
datasets 0
None public yet