Model comparison

Qwythos-9B-Claude-Mythos-5-1M vs Qwen3-VL-30B-A3B-Instruct

At Q4_K_M, Qwythos-9B-Claude-Mythos-5-1M is the smaller download — 11,516,776,704 bytes against 18,556,687,168. At long context the gap widens: Qwythos-9B-Claude-Mythos-5-1M's KV cache at 32K is 3.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Qwythos-9B-Claude-Mythos-5-1MQwen3-VL-30B-A3B-Instruct
Parameters9.4B31.1B
Architectureqwen35qwen3vlmoe
Layers3248
Native context1,048,576262,144
Mixture of expertsnoyes, 128 experts
Quantizations published1626
Smallest quantization5.38 GiB7.60 GiB
Q4_K_M10.73 GiB17.28 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextQwythos-9B-Claude-Mythos-5-1MQwen3-VL-30B-A3B-InstructRatio
4,0960.13 GiB0.38 GiB3.00×
8,1920.25 GiB0.75 GiB3.00×
16,3840.50 GiB1.50 GiB3.00×
32,7681.00 GiB3.00 GiB3.00×
65,5362.00 GiB6.00 GiB3.00×
131,0724.00 GiB12.00 GiB3.00×