Model comparison
Qwen3-VL-8B-Instruct-abliterated-v1 vs Llama-3.1-8B-Instruct
At Q4_K_M, Llama-3.1-8B-Instruct is the smaller download — 4,920,739,168 bytes against 5,027,785,792.
From the file· summed bytes, KV per layer
Side by side
| Qwen3-VL-8B-Instruct-abliterated-v1 | Llama-3.1-8B-Instruct | |
|---|---|---|
| Parameters | 8.8B | 8.0B |
| Architecture | qwen3vl | llama |
| Layers | — | 32 |
| Native context | — | 131,072 |
| Mixture of experts | no | no |
| Quantizations published | 48 | 45 |
| Smallest quantization | 1.97 GiB | 2.02 GiB |
| Q4_K_M | 4.68 GiB | 4.58 GiB |
| Licence | apache-2.0 | llama3.1 |
KV cache by context
the term that decides long-context viability
| Context | Qwen3-VL-8B-Instruct-abliterated-v1 | Llama-3.1-8B-Instruct | Ratio |
|---|---|---|---|
| 4,096 | — | 0.50 GiB | — |
| 8,192 | — | 1.00 GiB | — |
| 16,384 | — | 2.00 GiB | — |
| 32,768 | — | 4.00 GiB | — |
| 65,536 | — | 8.00 GiB | — |
| 131,072 | — | 16.00 GiB | — |