← AI Hashrate

Can B200 192GB SXM run Phi-4?

NVIDIA · 192 GB HBM3e · 8000 GB/s bandwidth · Microsoft · 14.7B · model ctx up to 16K

Yes — the B200 192GB SXM (192 GB HBM3e) can run Phi-4. At Q4 with the default 8K context it needs ≈ 12.03 GB VRAM (weights 8.09 GB + KV cache 2.94 GB + 1 GB runtime overhead), which fits within 192 GB. Decode at Q4/8K: ≈ 346.3 tok/s (estimated, batch 1). FP16 (≈ 33.34 GB) also fits, at ≈ 95.2 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

QuantContextVRAM neededFits?tok/s (decode)
Q44K10.56 GBYes346.3 est.
Q48K default12.03 GBYes346.3 est.
Q432K20.84 GBYes346.3 est.
FP164K31.87 GBYes95.2 est.
FP168K default33.34 GBYes95.2 est.
FP1632K42.16 GBYes95.2 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 neededB200 192GB SXM VRAM
Q48.09 GB2.94 GB1 GB12.03 GB192 GB
FP1629.4 GB2.94 GB1 GB33.34 GB192 GB

Other GPUs that run Phi-4

All GPUs for Phi-4 →

Other models for the B200 192GB SXM

All models on the B200 192GB SXM →