Model comparison
Qwen3.8-27B vs Yi-34B-200K-Llamafied
These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 3.7× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Qwen3.8-27B | Yi-34B-200K-Llamafied | |
|---|---|---|
| Parameters | 27.8B | 34.4B |
| Architecture | qwen35 | llama |
| Layers | 64 | 60 |
| Native context | 262,144 | 200,000 |
| Mixture of experts | no | no |
| Quantizations published | 22 | 21 |
| Smallest quantization | 8.39 GiB | 6.98 GiB |
| Q4_K_M | 15.66 GiB | — |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Qwen3.8-27B | Yi-34B-200K-Llamafied | Ratio |
|---|---|---|---|
| 4,096 | 0.25 GiB | 0.94 GiB | 3.75× |
| 8,192 | 0.50 GiB | 1.88 GiB | 3.75× |
| 16,384 | 1.00 GiB | 3.75 GiB | 3.75× |
| 32,768 | 2.00 GiB | 7.50 GiB | 3.75× |
| 65,536 | 4.00 GiB | 15.00 GiB | 3.75× |
| 131,072 | 8.00 GiB | 30.00 GiB | 3.75× |