Can I run Gemma-4-31B-Fable-5-Agent-Distill on a GeForce RTX 3050?
Not at these settings. No indexed quantization of Gemma-4-31B-Fable-5-Agent-Distill fits GeForce RTX 3050 at any context we compute, with q8_0 KV. The smallest shipped quantization is 12.82 GiB in weights alone, against 5.58 GiB usable. CPU offload can still run it, slowly.
Every quantization at every context
| Quant | Weights● | 4K◐ | 8K◐ | 16K◐ | 32K◐ | 64K◐ | 128K◐ |
|---|---|---|---|---|---|---|---|
| BF16 | 57.20 GiB | 59.0 | 59.4 | 60.0 | 61.4 | 64.0 | 69.3 |
| Q8_0 | 30.39 GiB | 32.2 | 32.6 | 33.2 | 34.6 | 37.2 | 42.5 |
| Q6_K | 23.47 GiB | 25.3 | 25.6 | 26.3 | 27.6 | 30.3 | 35.6 |
| Q5_K_M | 20.35 GiB | 22.2 | 22.5 | 23.2 | 24.5 | 27.2 | 32.5 |
| Q4_K_M | 17.40 GiB | 19.2 | 19.6 | 20.2 | 21.6 | 24.2 | 29.5 |
| IQ4_NL | 16.53 GiB | 18.4 | 18.7 | 19.4 | 20.7 | 23.3 | 28.7 |
| Q3_K_M | 14.24 GiB | 16.1 | 16.4 | 17.1 | 18.4 | 21.1 | 26.4 |
| Q3_K_S | 12.82 GiB | 14.7 | 15.0 | 15.6 | 17.0 | 19.6 | 24.9 |
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 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, most of this model's layers cache only a 1,024-token window rather than the full context.