Does openPangu-2.0-Flash fit in 12GB of VRAM?
Not at these settings. No indexed quantization of openPangu-2.0-Flash fits 12GB card at any context we compute, with q8_0 KV. The smallest shipped quantization is 40.73 GiB in weights alone, against 11.16 GiB usable. CPU offload can still run it, slowly.
Every quantization at every context
| Quant | Weights● | 4K◐ | 8K◐ | 16K◐ | 32K◐ | 64K◐ | 128K◐ |
|---|---|---|---|---|---|---|---|
| BF16 | 188.04 GiB | 189.0 | 189.1 | 189.3 | 189.7 | 190.6 | 192.3 |
| Q8_0 | 99.98 GiB | 100.9 | 101.0 | 101.2 | 101.6 | 102.5 | 104.2 |
| Q4_K_M | 56.71 GiB | 57.6 | 57.7 | 57.9 | 58.4 | 59.2 | 61.0 |
| Q3_K_M | 40.73 GiB | 41.6 | 41.8 | 42.0 | 42.4 | 43.3 | 45.0 |
Figures are GiB of total memory: weights plus KV cache plus compute buffer and backend overhead. Weights and KV are near-exact; the overhead term is modeled. Hover any cell for the breakdown.
Why there are no speeds on this page
A capacity is not a card. Whether a model fits depends only on memory, so every figure above holds for any 12GB accelerator. How fast it runs depends on memory bandwidth, which varies several-fold between cards of the same capacity — so putting a tokens-per-second number here would be inventing one. Pick a specific card from hardware and the speed column appears.
Why other calculators disagree
A parameters × bits ÷ 8 estimate ignores two things that dominate at long context. First, the weights themselves are not the nominal rate — quantizations are mixtures, so the real file is consistently larger than the label implies. Second, this model uses latent attention and allocates no V cache at all, so any formula reading num_key_value_heads overstates its cache by more than an order of magnitude.