Qwen3.6-35B-A3B-NVFP4 vs RnJ-1-Instruct-FP8

Specs, VRAM requirements and download trends — updated 6 August 2026.

nvidia · A

Params

18.7 B

Context

—

30d

11.3 M

apache-2.0 · commercial OK

Doradus-AI · B

Params

8.8 B

Context

32,768

30d

164 K

gemma licence · conditional

Specification comparison

Differences are highlighted; identical values are muted.

Specification	Qwen3.6-35B-A3B-NVFP4	RnJ-1-Instruct-FP8
Parameters	18.7 B	8.8 B
Architecture	Qwen3_5MoeForConditionalGeneration	Gemma3ForCausalLM
Context length	—	32,768
Licence	apache-2.0	gemma
Commercial use	Allowed	Conditional
Languages	—	1
Downloads 30d	11,298,952	163,514
Downloads all time	19.7 M	1.3 M
Quantizations on the Hub	SAFETENSORS	SAFETENSORS
Gated	No	No
First seen on the Hub	2026-05-27	2025-12-07

Daily snapshots, last 10 days (28 Jul – 6 Aug)

11.3 M164 K

Qwen3.6-35B-A3B-NVFP4 RnJ-1-Instruct-FP8

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

RnJ-1-Instruct-FP8 · fp16 15.0 GB / 24 GB

✅ Runs comfortably

Qwen3.6-35B-A3B-NVFP4 · fp16 29.1 GB / 24 GB

❌ Won’t fit