Model comparison
Qwen3.5-99B vs gpt-oss-20b
At Q4_K_M, gpt-oss-20b is the smaller download — 11,624,759,488 bytes against 60,194,980,672.
From the file· summed bytes, KV per layer
Side by side
| Qwen3.5-99B | gpt-oss-20b | |
|---|---|---|
| Parameters | 99.0B | 21.5B |
| Architecture | qwen35moe | gpt-oss |
| Layers | 48 | 24 |
| Native context | 262,144 | 131,072 |
| Mixture of experts | yes, 205 experts | yes, 32 experts |
| Quantizations published | 36 | 15 |
| Smallest quantization | 19.31 GiB | 10.68 GiB |
| Q4_K_M | 56.06 GiB | 10.83 GiB |
| Licence | other | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Qwen3.5-99B | gpt-oss-20b | Ratio |
|---|---|---|---|
| 4,096 | 0.09 GiB | 0.11 GiB | 1.19× |
| 8,192 | 0.19 GiB | 0.21 GiB | 1.09× |
| 16,384 | 0.38 GiB | 0.39 GiB | 1.05× |
| 32,768 | 0.75 GiB | 0.77 GiB | 1.02× |
| 65,536 | 1.50 GiB | 1.52 GiB | 1.01× |
| 131,072 | 3.00 GiB | 3.02 GiB | 1.01× |