APEX Quants (GGUF)
Collection
MoE models quantized with the APEX Quantization technique ( https://github.com/mudler/apex-quant ) • 23 items • Updated • 44
Nemotron-3-Super-120B-A12B APEX GGUF
APEX (Adaptive Precision for EXpert Models) quantizations of NVIDIA-Nemotron-3-Super-120B-A12B.
Brought to you by the LocalAI team | APEX Project | Technical Report
Benchmark Results
Benchmarks coming soon. For reference APEX benchmarks on the Qwen3.5-35B-A3B architecture, see mudler/Qwen3.5-35B-A3B-APEX-GGUF.
What is APEX?
APEX is a quantization strategy for Mixture-of-Experts (MoE) models. It classifies tensors by role (routed expert, shared expert, attention) and applies a layer-wise precision gradient -- edge layers get higher precision, middle layers get more aggressive compression. I-variants use diverse imatrix calibration (chat, code, reasoning, tool-calling, agentic traces, Wikipedia).
See the APEX project for full details, technical report, and scripts.
Architecture
- Model: NVIDIA-Nemotron-3-Super-120B-A12B (NemotronH)
- Layers: 88
- Type: Hybrid Mamba-2 / LatentMoE / Attention + Multi-Token Prediction (MTP)
- Experts: 512 routed + 1 shared (22 active per token)
- Total Parameters: 120B
- Active Parameters: ~12B per token
- APEX Config: 5+5 symmetric edge gradient across 88 layers
Run with LocalAI
local-ai run mudler/NVIDIA-Nemotron-3-Super-120B-A12B-APEX-GGUF@Nemotron-3-Super-120B-A12B-APEX-I-Balanced.gguf
Credits
APEX is brought to you by the LocalAI team. Developed through human-driven, AI-assisted research. Built on llama.cpp.
- Downloads last month
- 1,369
Hardware compatibility
Log In to add your hardware
We're not able to determine the quantization variants.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support