Qwen3.6-28B-REAP20-A3B-GGUF

Qwen3.6-28B-REAP20-A3B-GGUF

Authorbarozp
Parameters28B
QuantizationQ6_K
FormatGGUF
Licenseapache-2.0
File size21.64 GB
View on HuggingFace

Overview

A 20% expert-pruned variant of Qwen3.6-35B-A3B using the REAP method. 28B total parameters with 3B active, providing strong performance at reduced compute cost.

Strengths

  • BFCL (tool calling): 66.5% (400 problems)

Tested Hardware

Benchmarked by Infersec on:

  • NVIDIA RTX 5090 (64 GB RAM, 32 GB VRAM, linux)

Benchmark Runs

Performance and quality results across different hardware configurations