picksbycard
August 23, 2026No. 113
Home / Will it run? / Qwen3 30B-A3B / Radeon RX 9070
Radeon RX 9070 · 16GB GDDR6

Can the Radeon RX 9070 run Qwen3 30B-A3B?

Only with CPU offload at Q4_K_M.

Qwen3 30B-A3B is 30.5B parameters (MoE; all experts resident). At Q4_K_M the weights are about 17.0 GiB; with the 1.5 GiB runtime reserve that is 18.5 GiB against the card's 16GB, about 2.5 GiB more than the card has.

All three quants

16GB card
QuantWeightsWith reserveVerdictHeadroom
Q4_K_Mthe everyday quant17.0 GiB18.5 GiBOnly with CPU offloadshort 2.5 GiB
Q8_0near-lossless30.1 GiB31.6 GiBOnly with CPU offloadshort 15.6 GiB
FP16full weights56.8 GiB58.3 GiBDoes not fitshort 42.3 GiB

Verdict rules: fits comfortably when weights plus reserve sit within 85% of VRAM; tight when they fit with little left for context; CPU offload when they exceed VRAM by up to 2x (runs, slowly, with layers in system RAM); no beyond that. The KV cache per token depends on the model's architecture and is not modelled here.

Not this card. These fit it comfortably at Q4.

What the Radeon RX 9070 does run well → · Qwen3 30B-A3B on every card →

About the model

Qwen3 30B-A3B by Alibaba Qwen: 32,768-token native context, text, released 2025 under the Apache-2.0. Model card → · Hosted pricing on PicksByModel →

The Silicon Note

One short read on the GPU market, most mornings. No spam, one-click out.