Model comparison
Janus-Pro-7B vs FLUX.2-klein-4B
At Q4_K_M, FLUX.2-klein-4B is the smaller download — 2,497,280,416 bytes against 4,223,359,936.
From the file· summed bytes, KV per layer
Side by side
| Janus-Pro-7B | FLUX.2-klein-4B | |
|---|---|---|
| Parameters | 7.4B | 3.9B |
| Architecture | llama | flux |
| Layers | 30 | — |
| Native context | 16,384 | — |
| Mixture of experts | no | no |
| Quantizations published | 36 | 21 |
| Smallest quantization | 1.61 GiB | 1.70 GiB |
| Q4_K_M | 3.93 GiB | 2.33 GiB |
| Licence | mit | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Janus-Pro-7B | FLUX.2-klein-4B | Ratio |
|---|---|---|---|
| 4,096 | 1.88 GiB | — | — |
| 8,192 | 3.75 GiB | — | — |
| 16,384 | 7.50 GiB | — | — |
| 32,768 | 15.00 GiB | — | — |
| 65,536 | 30.00 GiB | — | — |
| 131,072 | 60.00 GiB | — | — |