← AI Hashrate

Can L40S 48GB run Devstral-Small-2507?

NVIDIA · 48 GB GDDR6 · 864 GB/s bandwidth · Mistral · 23.6B · model ctx up to 128K

Yes — the L40S 48GB (48 GB GDDR6) can run Devstral-Small-2507. At Q4 with the default 8K context it needs ≈ 18.7 GB VRAM (weights 12.98 GB + KV cache 4.72 GB + 1 GB runtime overhead), which fits within 48 GB. Decode at Q4/8K: ≈ 23.3 tok/s (estimated, batch 1). FP16 needs ≈ 52.92 GB — does not fit on this card. Estimates come from memory-bandwidth math; rows tagged measured override estimates. Relative ranking is more reliable than absolute tok/s. Methodology.

Fit & speed by quant and context

QuantContextVRAM neededFits?tok/s (decode)
Q44K16.34 GBYes23.3 est.
Q48K default18.7 GBYes23.3 est.
Q432K32.86 GBYes23.3 est.
FP164K50.56 GBNo16.5 est.
FP168K default52.92 GBNo15.1 est.
FP1632K67.08 GBNo9.4 est.

Fits = weights + KV(ctx) + 1 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).

VRAM breakdown at 8K context

QuantWeightsKV cacheOverheadTotal neededL40S 48GB VRAM
Q412.98 GB4.72 GB1 GB18.7 GB48 GB
FP1647.2 GB4.72 GB1 GB52.92 GB48 GB

Other GPUs that run Devstral-Small-2507

All GPUs for Devstral-Small-2507 →

Other models for the L40S 48GB

All models on the L40S 48GB →