đŦTechnical Deep Dive
Full Specifications [+]
Quick Commands
ollama run qwen3.5-27b-writer-v2-mlx-mixed-9.4bit huggingface-cli download thecluster/qwen3.5-27b-writer-v2-mlx-mixed-9.4bit âī¸ Nexus Index V2.0
đŦ Index Insight
FNI V2.0 for Qwen3.5 27b Writer V2 Mlx Mixed 9.4bit: Semantic (S:50), Authority (A:0), Popularity (P:11), Recency (R:98), Quality (Q:65).
Verification Authority
đ What's Next?
Technical Deep Dive
Qwen3.5-27B Writer V2
Quality: quantized (mixed quants per tensor, group size: 32, 9.450 bpw)
Most layers use 8-bit affine quantization with a group size 32; some layers are saved in bf16.
A writing & roleplay finetune of Qwen3.5 27B. The primary emphasis is on writing quality as it strongly generalizes across both domains.
Recommended Settings:
- Chatml template with
<think>\n\n</think>or{{char}}:prefill. Only non-thinking was trained, but thinking probably still works. - temperature =
0.7 - top_p =
0.95 - I do not recommend using high rep pen values like Qwen suggests for the base model. rep_pen =
1.05or a moderate dry setting should suffice.
Source
This model was converted to MLX format from ConicCat/Qwen3.5-27B-Writer-V2 using mlx-vlm version 0.4.4.
â ī¸ Incomplete Data
Some information about this model is not available. Use with Caution - Verify details from the original source before relying on this data.
View Original Source âđ Limitations & Considerations
- âĸ Benchmark scores may vary based on evaluation methodology and hardware configuration.
- âĸ VRAM requirements are estimates; actual usage depends on quantization and batch size.
- âĸ FNI scores are relative rankings and may change as new models are added.
- â License Unknown: Verify licensing terms before commercial use.
AI Summary: Based on Hugging Face metadata. Not a recommendation.
đĄī¸ Model Transparency Report
Technical metadata sourced from upstream repositories.
đ Identity & Source
- id
- hf-model--thecluster--qwen3.5-27b-writer-v2-mlx-mixed-9.4bit
- slug
- thecluster--qwen3.5-27b-writer-v2-mlx-mixed-9.4bit
- source
- huggingface
- author
- TheCluster
- license
- Apache-2.0
- tags
- mlx, safetensors, qwen3_5, 9bit, roleplay, conversational, mixed-precision, image-text-to-text, en, zh, ru, es, fr, it, ja, ko, af, de, ar, tr, is, pl, sw, sv, nl, he, id, uk, fa, pa, pt, ms, fi, el, base_model:coniccat/qwen3.5-27b-writer-v2, license:apache-2.0, 8-bit, region:us
âī¸ Technical Specs
- architecture
- Qwen3_5ForConditionalGeneration
- params billions
- 27.36
- context length
- 32,768
- pipeline tag
- image-text-to-text
- vram gb
- 23
- vram is estimated
- true
- vram formula
- VRAM â (params * 0.75) + 2GB (KV) + 0.5GB (OS)
đ Engagement & Metrics
- downloads
- 150
- stars
- 0
- forks
- 0
Data indexed from public sources. Updated daily.