Skip to content

Can Radeon AI PRO R9700 32G run Devstral Small 1.1 24B?

Comfortable: GGUF Q6_K needs 20.9 GB at 4K, batch 1; headroom 11.1 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_M15.7ComfortableVerified file
GGUF Q5_K_M18.2ComfortableVerified file
GGUF Q6_K20.9ComfortableVerified file
AWQ INT413.0ComfortableEstimated weights

Context capacity · GGUF Q6_K

ContextTotal GBVerdict
4K20.9Comfortable
16K22.9Comfortable
32K25.7Comfortable
128K42.2Does not fit

Generation speed

GGUF Q6_K: 18–24 tok/s · Derived range

Other model / GPU pairs