Skip to main content
Hybrid CPU+GPU

Best variant: Q8_0

CPU + GPU hybrid — not enough VRAM (12 GB < 27.6 GB min), but 64 GB RAM is sufficient. Expect significantly slower inference.

GPU VRAM
12 GB
Min VRAM (best fit)
27.6 GB
Recommended VRAM
31.2 GB
Estimated tok/s
~11

Share this matchup

Send this page so a friend can see if NVIDIA GeForce RTX 3080 Ti fits Devstral Small 2 24B.

Every Devstral Small 2 24B quantization on NVIDIA GeForce RTX 3080 Ti

Each row runs the compatibility engine against your VRAM, RAM, and the model's requirements.

QuantizationFile SizeMin VRAMRec VRAMContextVerdictEstimated tok/s
Q4_K_M12 GB13.8 GB15.6 GB8K / 8KHybrid CPU+GPU~19
Q5_K_M15 GB17.3 GB19.5 GB8K / 8KHybrid CPU+GPU~17
Q8_0Best fit24 GB27.6 GB31.2 GB8K / 8KHybrid CPU+GPU~11
FP1648 GB55.2 GB62.4 GB8K / 8KCan't Run

Upgrade options that fit Devstral Small 2 24B better

Cheapest fit

NVIDIA GeForce RTX 5090

32 GB VRAM · ~71.1 tok/s

Best value

Apple M1 Pro (10-core GPU)

32 GB VRAM · ~7.9 tok/s

Best performance

Apple M4 Ultra

256 GB VRAM · ~43.3 tok/s

Rent GPU instead of buying one

If local fit is weak, cloud GPU gets you running today without hardware upgrade.

All hardware for Devstral Small 2 24BBest GPU for Devstral Small 2 24BModels that fit NVIDIA GeForce RTX 3080 TiFull model detailsBrowse all models