Model comparison
Qwen3-Coder-30B-A3B-Instruct vs Omega-Evolution-26B-A4B-v3.0
At Q4_K_M, Qwen3-Coder-30B-A3B-Instruct is the smaller download — 18,556,689,568 bytes against 51,258,890,720.
From the file· summed bytes, KV per layer
Side by side
| Qwen3-Coder-30B-A3B-Instruct | Omega-Evolution-26B-A4B-v3.0 | |
|---|---|---|
| Parameters | 30.5B | 26.5B |
| Architecture | qwen3moe | gemma4 |
| Layers | 48 | — |
| Native context | 262,144 | — |
| Mixture of experts | yes, 128 experts | no |
| Quantizations published | 46 | 5 |
| Smallest quantization | 7.46 GiB | 13.10 GiB |
| Q4_K_M | 17.28 GiB | 47.74 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Qwen3-Coder-30B-A3B-Instruct | Omega-Evolution-26B-A4B-v3.0 | Ratio |
|---|---|---|---|
| 4,096 | 0.38 GiB | — | — |
| 8,192 | 0.75 GiB | — | — |
| 16,384 | 1.50 GiB | — | — |
| 32,768 | 3.00 GiB | — | — |
| 65,536 | 6.00 GiB | — | — |
| 131,072 | 12.00 GiB | — | — |