Qwen 3.5 27B GGUF — Quantized by BatiAI
🆕 Newer model available — Qwen3.8-27B GGUF Qwen's 2026-08 release. Same 27B class, six quants from 10 GB, vision included, Korean-verified.
ollama run batiai/qwen3.8-27b:iq4
IQ4_XS quantization of Qwen/Qwen3.5-27B for on-device AI on Mac. Built and verified by BatiAI for BatiFlow.
Quick Start
ollama pull batiai/qwen3.5-27b:iq4
Available Quantizations
| Quant | Size | VRAM | M4 Max (128GB) | Recommended For |
|---|---|---|---|---|
| IQ4_XS | 14GB | 28GB | 17.0 t/s | 32GB+ Mac |
Benchmarks — M4 Max (128GB)
| Metric | IQ4_XS |
|---|---|
| Token generation | 17.0 t/s |
| Korean | ✅ |
| Tool call JSON | ✅ |
| VRAM | 28 GB |
vs Other Qwen 3.5 Models
| Model | Size | VRAM | Speed | Min Mac |
|---|---|---|---|---|
| batiai/qwen3.5-9b:q4 | 5.2GB | ~8GB | 12.5 t/s | 16GB |
| batiai/qwen3.5-27b:iq4 | 14GB | 28GB | 17.0 t/s | 32GB |
| batiai/qwen3.5-35b:iq4 | 17GB | 23GB | 26.6 t/s | 36GB |
For 36GB+ Mac, consider batiai/qwen3.5-35b — MoE architecture, faster and less VRAM.
Technical Details
- Original Model: Qwen/Qwen3.5-27B
- Architecture: Hybrid (Gated DeltaNet + GQA + MoE)
- Context Window: 262K tokens
- License: Apache 2.0
- Quantized with: llama.cpp (build 400ac8e)
About BatiFlow
BatiFlow — free, on-device AI automation for Mac. 5MB app, 100% local, unlimited.
License
Quantized from Qwen/Qwen3.5-27B. License: Apache 2.0.
Benchmarks
| Machine | Quant | Cold start | Prompt eval | Token gen | Tested |
|---|---|---|---|---|---|
| MacBook Pro M4 Max 128GB | IQ4_XS | 4.827s | 82.37 t/s | 11.71 t/s | 2026-05-03 |