NVIDIA · 80 GB HBM3 · 3350 GB/s bandwidth ·
Mistral · 23.6B · model ctx up to 128K
Yes — the H100 80GB SXM5 (80 GB HBM3) 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 80 GB. Decode at Q4/8K: ≈ 90.3 tok/s (estimated, batch 1). FP16 (≈ 52.92 GB) also fits, at ≈ 24.8 tok/s (estimated). 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
Quant
Context
VRAM needed
Fits?
tok/s (decode)
Q4
4K
16.34 GB
Yes
90.3est.
Q4
8K default
18.7 GB
Yes
90.3est.
Q4
32K
32.86 GB
Yes
90.3est.
FP16
4K
50.56 GB
Yes
24.8est.
FP16
8K default
52.92 GB
Yes
24.8est.
FP16
32K
67.08 GB
Yes
24.8est.
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).