Model comparison
Qwen3-42B-A3B-2507-Thinking-Abliterated-uncensored-TOTAL-RECALL-v2-Medium-MASTER-CODER vs Qwen3-Coder-30B-A3B-Instruct
At Q4_K_M, Qwen3-Coder-30B-A3B-Instruct is the smaller download — 18,556,689,568 bytes against 25,703,283,040. At long context the gap widens: Qwen3-Coder-30B-A3B-Instruct's KV cache at 32K is 1.4× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Qwen3-42B-A3B-2507-Thinking-Abliterated-uncensored-TOTAL-RECALL-v2-Medium-MASTER-CODER | Qwen3-Coder-30B-A3B-Instruct | |
|---|---|---|
| Parameters | 42.4B | 30.5B |
| Architecture | qwen3moe | qwen3moe |
| Layers | 67 | 48 |
| Native context | 262,144 | 262,144 |
| Mixture of experts | yes, 128 experts | yes, 128 experts |
| Quantizations published | 34 | 46 |
| Smallest quantization | 8.21 GiB | 7.46 GiB |
| Q4_K_M | 23.94 GiB | 17.28 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Qwen3-42B-A3B-2507-Thinking-Abliterated-uncensored-TOTAL-RECALL-v2-Medium-MASTER-CODER | Qwen3-Coder-30B-A3B-Instruct | Ratio |
|---|---|---|---|
| 4,096 | 0.52 GiB | 0.38 GiB | 1.40× |
| 8,192 | 1.05 GiB | 0.75 GiB | 1.40× |
| 16,384 | 2.09 GiB | 1.50 GiB | 1.40× |
| 32,768 | 4.19 GiB | 3.00 GiB | 1.40× |
| 65,536 | 8.38 GiB | 6.00 GiB | 1.40× |
| 131,072 | 16.75 GiB | 12.00 GiB | 1.40× |