Skip to content

Can Radeon RX 7900 XT run MiniCPM-V 4.7 35B-A3B?

Tight: GGUF Q4_K_S needs 20.9 GB at 4K, batch 1; headroom -0.9 GB.

Text input, FP16 KV cache, one card. Unified-memory capacity uses the GPU guide’s usable limit; negative headroom may require host offload. Speed source settings may differ from this 4K estimate.

Measured runs

No measured run for this pair. Sizing is an estimate, not a runtime test.

Quant memory at 4K

QuantTotal GBVerdictWeight sizing
GGUF Q4_K_S20.9TightVerified file
GGUF Q4_K_M22.2Does not fitVerified file
GGUF Q5_K_M25.9Does not fitVerified file
GGUF Q8_038.6Does not fitVerified file

Context capacity · GGUF Q4_K_S

ContextTotal GBVerdict
4K20.9Tight
16K21.1Does not fit
32K21.5Does not fit
128K23.6Does not fit

Generation speed

GGUF Q4_K_S: 489 tok/s · Bandwidth ceiling (not measured)

Other model / GPU pairs