← AI Hashrate

Can Gaudi 3 128GB run Qwen3.5-2B?

Intel · 128 GB HBM2e · 3276 GB/s bandwidth · Alibaba · 2.0B · model ctx up to 256K

Yes — the Gaudi 3 128GB (128 GB HBM2e) can run Qwen3.5-2B. At Q4 with the default 8K context it needs ≈ 2.5 GB VRAM (weights 1.1 GB + KV cache 0.4 GB + 1 GB runtime overhead), which fits within 128 GB. Decode at Q4/8K: ≈ 1042.4 tok/s (estimated, batch 1). FP16 (≈ 5.4 GB) also fits, at ≈ 286.6 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)
Q44K2.3 GBYes1042.4 est.
Q48K default2.5 GBYes1042.4 est.
Q432K3.7 GBYes1042.4 est.
FP164K5.2 GBYes286.6 est.
FP168K default5.4 GBYes286.6 est.
FP1632K6.6 GBYes286.6 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 neededGaudi 3 128GB VRAM
Q41.1 GB0.4 GB1 GB2.5 GB128 GB
FP164.0 GB0.4 GB1 GB5.4 GB128 GB

Other GPUs that run Qwen3.5-2B

All GPUs for Qwen3.5-2B →

Other models for the Gaudi 3 128GB

All models on the Gaudi 3 128GB →