Qwen3.6-35B-A3B-NVFP4 vs Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP
Specs, VRAM requirements and download trends — updated 25 October 2026.
nvidia · A · Open weights
Params
18.7 B
Context
—
30d
4.1 M
apache-2.0 · commercial OK
sakamakismile · B · Open weights
Params
27.0 B
Context
—
30d
195 K
apache-2.0 · commercial OK
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | Qwen3.6-35B-A3B-NVFP4 | Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP |
|---|---|---|
| Parameters | 18.7 B | 27.0 B |
| Architecture | Qwen3_5MoeForConditionalGeneration | Qwen3_5ForConditionalGeneration |
| Context length | — | — |
| Licence | apache-2.0 | apache-2.0 |
| Commercial use | Allowed | Allowed |
| Languages | — | — |
| Downloads 30d | 4,142,396 | 195,002 |
| Downloads all time | 39.6 M | 562 K |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | No | No |
| First seen on the Hub | 2026-05-27 | 2026-04-25 |
Download trend
Daily snapshots, last 90 days (28 Jul – 25 Oct)
12.6 M195 K
Qwen3.6-35B-A3B-NVFP4
Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP · fp16
27.2 GB / 24 GB
Qwen3.6-35B-A3B-NVFP4 · fp16
29.1 GB / 24 GB