How much VRAM does gpt-oss-20b-BF16 need?

Its published quantizations range from 12.03 GiB to 38.99 GiB in weights alone, before KV cache and runtime overhead.

Can I run gpt-oss-20b-BF16 on a Radeon RX 6500 XT?

Q: Can I run gpt-oss-20b-BF16 on a Radeon RX 6500 XT?

No. No published quantization of gpt-oss-20b-BF16 fits a Radeon RX 6500 XT at any context with q4_0 KV cache.

Not at these settings. No indexed quantization of gpt-oss-20b-BF16 fits Radeon RX 6500 XT at any context we compute, with q4_0 KV. The smallest shipped quantization is 12.03 GiB in weights alone, against 3.72 GiB usable. CPU offload can still run it, slowly.

From the file· weights summed from filesFrom the file· KV computed per layerPredicted· speed and compute buffer

KV cache dtypef16 q8_0 q4_0— 4 GB card, 3.72 GiB usable after overhead

Every quantization at every context

total memory required; green fits, red does not

Quant	Weights●	4K◐	8K◐	16K◐	32K◐	64K◐	128K◐
F16	38.99 GiB	39.9	39.9	40.0	40.1	40.3	40.7
Q8_0	20.73 GiB	21.7	21.7	21.7	21.8	22.0	22.5
Q5_K_M	15.73 GiB	16.7	16.7	16.7	16.8	17.0	17.5
Q4_K_M	14.72 GiB	15.6	15.7	15.7	15.8	16.0	16.5
Q3_K_M	12.03 GiB	12.9	13.0	13.0	13.1	13.3	13.8

●From the file◐Predictedwhat these mean

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 128-token window rather than the full context.