Model comparison

rnj-1-instruct vs Qwen3-VL-8B-Instruct-abliterated-v1

At Q4_K_M, Qwen3-VL-8B-Instruct-abliterated-v1 is the smaller download — 5,027,785,792 bytes against 5,113,914,368.

From the file· summed bytes, KV per layer

Side by side

rnj-1-instructQwen3-VL-8B-Instruct-abliterated-v1
Parameters8.3B8.8B
Architecturegemma3qwen3vl
Layers32
Native context32,768
Mixture of expertsnono
Quantizations published4648
Smallest quantization1.98 GiB1.97 GiB
Q4_K_M4.76 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextrnj-1-instructQwen3-VL-8B-Instruct-abliterated-v1Ratio
4,0960.50 GiB
8,1921.00 GiB
16,3842.00 GiB
32,7684.00 GiB
65,5368.00 GiB
131,07216.00 GiB
rnj-1-instruct vs Qwen3-VL-8B-Instruct-abliterated-v1 — size, memory and hardware fit — ossmodeldb