Qwen3.8-27B-MLX-4bit vs DeepSeek-OCR-2

Specs, VRAM requirements and download trends — updated 21 September 2026.

lmstudio-community · A
Params
4.7 B
Context
30d
4.7 M
apache-2.0 · commercial OK
deepseek-ai · B
Params
3.4 B
Context
8,192
30d
900 K
apache-2.0 · commercial OK

Specification comparison

Differences are highlighted; identical values are muted.

Specification Qwen3.8-27B-MLX-4bit DeepSeek-OCR-2
Parameters 4.7 B 3.4 B
Architecture Qwen3_5ForConditionalGeneration DeepseekOCR2ForCausalLM
Context length 8,192
Licence apache-2.0 apache-2.0
Commercial use Allowed Allowed
Languages 1
Downloads 30d 4,696,719 899,794
Downloads all time 5.8 M 14.1 M
Quantizations on the Hub SAFETENSORS SAFETENSORS
Gated No No
First seen on the Hub 2026-08-14 2026-01-27

Download trend

Daily snapshots, last 31 days (22 Aug – 21 Sep)

5.0 M889 K
Qwen3.8-27B-MLX-4bit DeepSeek-OCR-2 A overtook B on 25 Aug

VRAM side by side

On RTX 4090 · 24 GB · 8K context unless noted

DeepSeek-OCR-2 · fp16 8.5 GB / 24 GB
✅ Runs comfortably
Qwen3.8-27B-MLX-4bit · fp16 18.9 GB / 24 GB
✅ Runs comfortably

Adjacent comparisons