Mistral on AMD RDNA3

Can I run Mixtral 8x7B on an Radeon RX 7900 XTX?

Not at Q4_K_M — it needs 28.2 GB against 21.8 GB available. Drop to IQ3_XXS and it fits, at about 86.4 tokens per second.

Won't fit

28.2 GB of 21.8 GB · 130%
029 GB
Weights 26.4 GB
KV cache 1.0 GB
Runtime overhead 0.8 GB
Over the limit 6.4 GB

Short by 6.4 GB. You can run it with 24 of 32 layers on the Radeon RX 7900 XTX and the rest in system RAM, at roughly 15.0 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.

Generation15.0tok/s
Prompt processing2002tok/s
Max context0tokens
KV per 1K tokens0GB

Every quantisation of Mixtral 8x7B on a Radeon RX 7900 XTX

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 87.0 GB 88.8 GB Won't fit 1.85
INT8 / W8A8 8.50 46.2 GB 48.0 GB Won't fit 4.37
Q8_0 (GGUF) 8.50 46.2 GB 48.0 GB Won't fit 4.37
FP8 (E4M3) 8.00 43.5 GB 45.3 GB Won't fit 4.85
Q6_K 6.56 35.7 GB 37.5 GB Won't fit 6.85
Q5_K_M 5.67 30.8 GB 32.6 GB Won't fit 9.46
Q5_K_S 5.52 30.0 GB 31.8 GB Won't fit 10.4
Q4_K_M 4.85 26.4 GB 28.2 GB Won't fit 15.0
Q4_K_S 4.58 24.9 GB 26.7 GB Won't fit 17.4
Q4_0 4.55 24.8 GB 26.6 GB Won't fit 17.5
AWQ 4-bit 4.25 23.5 GB 25.3 GB Won't fit 23.1
GPTQ 4-bit 4.25 23.5 GB 25.3 GB Won't fit 23.1
MXFP4 4.25 23.5 GB 25.3 GB Won't fit 23.1
IQ4_XS 4.25 23.1 GB 25.0 GB Won't fit 23.4
Q3_K_M 3.91 21.3 GB 23.1 GB Won't fit 33.9
IQ3_M 3.70 20.2 GB 22.0 GB Won't fit 6K 54.3
IQ3_XXS 3.06 16.7 GB 18.5 GB Runs comfortably 32K 86.4
Q2_K 2.63 14.4 GB 16.2 GB Runs comfortably 32K 98.0
IQ2_XXS 2.06 11.3 GB 13.1 GB Runs comfortably 32K 119
IQ1_M 1.75 9.6 GB 11.4 GB Runs comfortably 32K 135

Also worth checking