Model comparison

Qwen3-32B-qweh2 vs Qwen-AgentWorld-35B-A3B

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: Qwen-AgentWorld-35B-A3B's KV cache at 32K is 12.8× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Qwen3-32B-qweh2Qwen-AgentWorld-35B-A3B
Parameters32.8B34.7B
Architectureqwen3qwen35moe
Layers6440
Native context40,960262,144
Mixture of expertsnoyes, 256 experts
Quantizations published115
Smallest quantization21.62 GiB10.71 GiB
Q4_K_M
Licencewtfplapache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3-32B-qweh2Qwen-AgentWorld-35B-A3BRatio
4,0961.00 GiB0.08 GiB12.79×
8,1922.00 GiB0.16 GiB12.80×
16,3844.00 GiB0.31 GiB12.80×
32,7688.00 GiB0.63 GiB12.80×
65,53616.00 GiB1.25 GiB12.80×
131,07232.00 GiB2.50 GiB12.80×