Model comparison
OpenCoder-1.5B-Instruct vs Qwen3-8B
At Q4_K_M, OpenCoder-1.5B-Instruct is the smaller download — 1,417,775,200 bytes against 5,027,783,488. At long context the gap widens: Qwen3-8B's KV cache at 32K is 1.5× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| OpenCoder-1.5B-Instruct | Qwen3-8B | |
|---|---|---|
| Parameters | 1.9B | 8.2B |
| Architecture | llama | qwen3 |
| Layers | 24 | 36 |
| Native context | 4,096 | 40,960 |
| Mixture of experts | no | no |
| Quantizations published | 4 | 51 |
| Smallest quantization | 1.18 GiB | 2.12 GiB |
| Q4_K_M | 1.32 GiB | 4.68 GiB |
| Licence | other | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | OpenCoder-1.5B-Instruct | Qwen3-8B | Ratio |
|---|---|---|---|
| 4,096 | 0.82 GiB | 0.56 GiB | 1.46× |
| 8,192 | 1.64 GiB | 1.13 GiB | 1.46× |
| 16,384 | 3.28 GiB | 2.25 GiB | 1.46× |
| 32,768 | 6.56 GiB | 4.50 GiB | 1.46× |
| 65,536 | 13.13 GiB | 9.00 GiB | 1.46× |
| 131,072 | 26.25 GiB | 18.00 GiB | 1.46× |