Model comparison

Mistral-Nemo-Instruct-2407-12B-Thinking-M-Claude-Opus-High-Reasoning vs Qwen3-8B

These two publish different quantization sets; the table below has the exact sizes.

From the file· summed bytes, KV per layer

Side by side

Mistral-Nemo-Instruct-2407-12B-Thinking-M-Claude-Opus-High-ReasoningQwen3-8B
Parameters12.2B8.2B
Architecturellamaqwen3
Layers4036
Native context1,024,00040,960
Mixture of expertsnono
Quantizations published2451
Smallest quantization2.79 GiB2.12 GiB
Q4_K_M—4.68 GiB
Licence—apache-2.0

KV cache by context

the term that decides long-context viability
ContextMistral-Nemo-Instruct-2407-12B-Thinking-M-Claude-Opus-High-ReasoningQwen3-8BRatio
4,0960.63 GiB0.56 GiB1.11×
8,1921.25 GiB1.13 GiB1.11×
16,3842.50 GiB2.25 GiB1.11×
32,7685.00 GiB4.50 GiB1.11×
65,53610.00 GiB9.00 GiB1.11×
131,07220.00 GiB18.00 GiB1.11×