DeskGPU
English

Can AMD Radeon RX 7900 XTX run Qwen3-Coder-30B-A3B-Instruct?

This page was written and translated automatically by AI and checked by software against its sources, with no human editing. Specifications come from the sources listed; memory requirements and speeds are estimates calculated with published formulas. How we use AI

Fits By llama.cpp’s default load check, Qwen3-Coder-30B-A3B-Instruct fits on AMD Radeon RX 7900 XTX at Q4_K_M with a context of 8,192 tokens. One card needs 20.1 GiB available for “Fits”.

At Q4_K_M, the longest context that fits is 34,816 tokens.

“Fits” means llama.cpp's default --fit load check would place all layers on the GPU: our estimate of llama.cpp's buffers, plus its error margin, plus llama.cpp's 1 GiB per-card reserve, is within the memory counted as available. It does not promise that peak memory while generating fits. How results are decided

Values updated Oct 19, 2026 · Calculator

Run profile: llama.cpp, ROCm back end on Linux (AMD discrete GPU), all layers on the GPU, Flash Attention on, micro-batch (-ub) 512, F16 KV cache, context (-c) as shown; “Fits” is judged with llama-server’s default 4 parallel sequences (-np auto, unified KV cache) and “Doesn’t fit” with 1 sequence; idle GPU with no other programs using its memory.

Memory counted as available for “Fits” (C_low): 22.75 GiB per card = 24 GiB nominal − 16 MiB (nominal minus the total that llama.cpp reported for this GPU in public logs) − 1,216 MiB (the largest start-up and runtime use seen in this profile’s public logs), rounded down to a multiple of 64 MiB.

These results apply to the configuration seen in public logs, where llama.cpp reported 24,560 MiB of total memory for this GPU. Settings that reduce usable memory (for example ECC, where the card supports it) are not covered.

Memory used for “Doesn’t fit” (C_high): the full 24 GiB per card. Run profiles and their measurements

Memory needed by quantization and context

Available memory needed for “Fits” (X, GiB per card) and the result on AMD Radeon RX 7900 XTX.
Quantization4,0968,19232,768131,072
BF16 61.0 GiB Doesn't fit61.4 GiB Doesn't fit63.8 GiB Doesn't fit73.5 GiB Doesn't fit
Q8_0 33.2 GiB Doesn't fit33.6 GiB Doesn't fit36.0 GiB Doesn't fit45.7 GiB Doesn't fit
Q6_K 26.1 GiB Tight26.5 GiB Tight28.9 GiB Doesn't fit38.5 GiB Doesn't fit
Q5_K_M 22.8 GiB Tight23.2 GiB Tight25.6 GiB Tight35.3 GiB Doesn't fit
Q4_K_M 19.7 GiB Fits20.1 GiB Fits22.5 GiB Fits32.2 GiB Doesn't fit
Q4_0 18.6 GiB Fits19.0 GiB Fits21.4 GiB Fits31.1 GiB Doesn't fit
IQ4_XS 17.6 GiB Fits18.0 GiB Fits20.4 GiB Fits30.1 GiB Doesn't fit
Q3_K_M 16.0 GiB Fits16.4 GiB Fits18.8 GiB Fits28.5 GiB Tight
Q2_K 12.6 GiB Fits13.0 GiB Fits15.5 GiB Fits25.1 GiB Tight

No speed ceiling is given: the memory bandwidth of AMD Radeon RX 7900 XTX is not confirmed by two independent open sources.

Alternatives

Other 24 GB GPUs where Qwen3-Coder-30B-A3B-Instruct fits

Sources