Estimates from memory-bandwidth formula; rows marked "measured" override. Methodology.
Fit & speed by quant and context
Quant
Context
VRAM needed
Fits?
tok/s (decode)
Q4
4K
3.8 GB
Yes
18.9±est.
Q4
8K default
4.37 GB
Yes
18.9±est.
Q4
32K
7.74 GB
No
15.9±est.
FP16
4K
9.56 GB
No
2.8±est.
FP16
8K default
10.12 GB
No
2.5±est.
FP16
32K
13.5 GB
No
1.4±est.
Fits = weights + KV(ctx) + 0.5 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).