← AI Hashrate

Can MI300X 192GB run gemma-4-31B-it?

AMD · 192 GB HBM3 · 5300 GB/s bandwidth · Google · 30.7B · model ctx up to 256K

Yes — the MI300X 192GB (192 GB HBM3) can run gemma-4-31B-it. At Q4 with the default 8K context it needs ≈ 24.03 GB VRAM (weights 16.89 GB + KV cache 6.14 GB + 1 GB runtime overhead), which fits within 192 GB. Decode at Q4/8K: ≈ 109.9 tok/s (estimated, batch 1). FP16 (≈ 68.54 GB) also fits, at ≈ 30.2 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)
Q44K20.96 GBYes109.9 est.
Q48K default24.03 GBYes109.9 est.
Q432K42.45 GBYes109.9 est.
FP164K65.47 GBYes30.2 est.
FP168K default68.54 GBYes30.2 est.
FP1632K86.96 GBYes30.2 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 neededMI300X 192GB VRAM
Q416.89 GB6.14 GB1 GB24.03 GB192 GB
FP1661.4 GB6.14 GB1 GB68.54 GB192 GB

Other GPUs that run gemma-4-31B-it

All GPUs for gemma-4-31B-it →

Other models for the MI300X 192GB

All models on the MI300X 192GB →