llmrun

Qwen 3.5 Models — Hardware Requirements

34 Qwen 3.5 models from Alibaba and the community, from the smallest that runs in 0.6 GB of VRAM up to 403.4B parameters. Every row links to full quantization tables and GPU compatibility.

All Qwen 3.5 Models by Size

ModelParamsContext
Qwen3.5 4B DFlash634M262K
Josiefied Qwen3.5 0.8B Gabliterated V1853M262K
Qwen3.5 0.8B873M262K
Qwen3.5 0.8B Base873M262K
Qwen3.5 9B DFlash1.3B262K
Qwen3.5 2B Text Only1.9B262K
Qwen3.5 2B2.3B262K
Qwen3.5 2B Base2.3B262K
Qwen3.5 2B Claude 4.6 Opus Reasoning Distilled2.3B262K
Qwen3.5 4B Super Coder4B262K
Qwen3.5 4B PTBR4B—
Qwen3.5 4B Safety Thinking4.2B262K
Qwen3.5 4B Claude Opus 4.6 Distilled Heretic4.5B262K
Qwen3.5 4B EmperoAI Qwen3.8 Distill Heretic Abliterated4.5B262K
Qwen3.5 4B Base4.7B262K
Qwen3.5 4B4.7B262K
Qwen3.5 4B Claude 4.6 Opus Reasoning Distilled4.7B262K
Qwen3.5 4B MiniFantasy MTP4.7B262K
Qwen3.5 9B Abliterated9.0B262K
Qwen3.5 9B Uncensored9B—
Qwen3.5 9B Humanize DPO Round29B—
GrepSeek Qwen3.5 9B GRPO9.4B262K
Qwen3.5 9B9.7B262K
Qwen3.5 9B The Defiant Fable Uncensored Heretic NEO IMATRIX MAX MTP9.7B262K
Qwen3.5 9B Claude 4.6 Opus Reasoning Distilled9.7B262K
Qwen3.5 9B Base9.7B262K
Qwen3.5 9B Gemini 3.1 Pro Reasoning Distill9.7B262K
Qwen 3.5 Sql 27B Distill 2B27B262K
Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Heretic v227.4B262K
Qwen3.5 27B27.8B262K
Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled27.8B262K
Qwen3.5 35B A3B DFlash35B262K
PrunedHub Qwen3.5 35B A3B 80pct35B—
Qwen3.5 35B A3B Base36.0B262K
Qwen3.5 35B A3B Claude 4.6 Opus Reasoning Distilled36.0B262K
Qwen3.5 35B A3B36.0B262K
Qwen3.5 122B A10B125.1B262K
Qwen3.5 397B A17B403.4B262K

How Qwen 3.5 Compares — Benchmark Rating

Qwen3.5 397B A17B is the highest-rated Qwen 3.5 model with an overall benchmark rating of 66.1/100 — #2 among 78 open models. The top proprietary model, Claude Fable 5 (high), scores 87.8. Click a model to see its full benchmark breakdown.

Claude Fable 5 (high) · proprietary87.8
Claude Opus 4.7 · proprietary86.4
Claude Fable 5 Max · proprietary82.4
GPT 6 Astra (high) · proprietary82.4
GPT 6 Astra Max · proprietary82.4
Kimi K366.8
Qwen3.5 397B A17B66.1
GLM 565.9
DeepSeek V4.1 Flash65.8
DeepSeek R1 052863.1
DeepSeek V4 Pro 081362.6
Qwen3.5 27B58.8
Qwen3.5 35B A3B54.7
Qwen3.5 122B A10B52.2
Composite of normalized public benchmark scores (methodology) · ■ Qwen 3.5 · ■ other models

Frequently Asked Questions

How much VRAM do I need to run a Qwen 3.5 model?
The smallest Qwen 3.5 model, Qwen3.5 0.8B, runs from 0.6 GB of VRAM at an aggressive quantization. Larger family members need proportionally more — see the table above for every model.
Which Qwen 3.5 models can I run on a 16 GB GPU?
30 of 38 Qwen 3.5 models fit in 16 GB of VRAM at some quantization, including Qwen3.5 9B, Qwen3.5 9B The Defiant Fable Uncensored Heretic NEO IMATRIX MAX MTP, Qwen3.5 2B.
What is the most popular Qwen 3.5 model to run locally?
Qwen3.5 9B is the most downloaded Qwen 3.5 model in local-friendly quantized formats. It runs from 3.2 GB of VRAM.
How do Qwen 3.5 models score on benchmarks?
Qwen3.5 397B A17B leads the family with an overall benchmark rating of 66.1/100, ranking #2 among 78 open models, while the top proprietary model, Claude Fable 5 (high), scores 87.8. See the comparison chart above for the full standings.