MINT quantized Nemotron-3-Super-120B — hybrid Mamba-MoE-Attention (MLX & GGUF)
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 — bringing 400B+ parameter models to a single machine.
Website · Research · GitHub · MINT Space
Research
Read the full paper: MINT: Compute-Optimal Data-Free Mixed-Precision Quantization for LLMs
- Papers
- Benchmark Results
- Key Findings
- Technical Deep Dives
Browse all research at baa.ai/articles
spaces 6
Running
1
MINT
🌿
Quantize LLMs to fit a specific memory budget without data
Running
Per-Expert Mixed-Precision Quantization for 512-Expert MoE Models
🔬
Quantize large MoE models to mixed precision without data
Running
1
SWAN: Data-Free Mixed-Precision Quantization
🦢
Quantize LLMs without data using per‑tensor mixed precision
Running
2
SAKD: SWAN-Guided Knowledge Distillation
📖
Generate quantization‑ready student models via guided distillation
Running
1
Sensitivity-Aware Training (SAT)
📄
Train LLMs to be quantization‑ready with sensitivity‑aware methods
models 36
baa-ai/Llama-4-Maverick-17B-128E-Instruct-MINT-407GB-GGUF
Text Generation • 401B • Updated • 43
baa-ai/Llama-4-Maverick-17B-128E-Instruct-SWAN-4bit-MLX
51B • Updated • 247
baa-ai/Llama-4-Scout-17B-16E-Instruct-MINT-117GB-MLX
Text Generation • 108B • Updated • 411
baa-ai/Llama-4-Maverick-17B-128E-Instruct-MINT-407GB-MLX
Text Generation • 401B • Updated • 252
baa-ai/Llama-4-Scout-17B-16E-Instruct-SWAN-4bit-MLX
18B • Updated • 308
baa-ai/Llama-4-Scout-17B-16E-Instruct-MINT-117GB-GGUF
Text Generation • 108B • Updated • 325
baa-ai/Nemotron-3-Super-120B-A12B-MINT-GGUF
121B • Updated • 112
baa-ai/Nemotron-3-Super-120B-A12B-MINT-MLX
121B • Updated • 181
baa-ai/Qwen3-30B-A3B-SWAN-5bit-MLX
31B • Updated • 403 • 1
baa-ai/Qwen3-30B-A3B-MINT-4bit-MLX
31B • Updated • 125
datasets 0
None public yet