← AI Hashrate 中文

Can Arc Pro B70 32GB run Qwen3.6-35B-A3B?

Intel · 32 GB GDDR6 · 608 GB/s bandwidth · Alibaba · 36.0B (active 3.0B) · MoE · model ctx up to 256K

BB · GoodComposite score: VRAM headroom × 0.5 + tok/s speed × 0.5. Q4 @ 8K context.
ONEAPI

Yes — the Arc Pro B70 32GB (32 GB GDDR6) can run Qwen3.6-35B-A3B. At Q4 with the default 8K context it needs ≈ 20.88 GB VRAM (weights 19.8 GB + KV cache 0.08 GB + 1 GB runtime overhead), which fits within 32 GB. Decode at Q4/8K: ≈ 129.0 tok/s (estimated, batch 1). FP16 needs ≈ 73.08 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)
Q44K20.84 GBYes129.0 est.
Q48K default20.88 GBYes129.0 est.
Q432K21.1 GBYes129.0 est.
FP164K73.04 GBNo19.4 est.
FP168K default73.08 GBNo19.4 est.
FP1632K73.3 GBNo19.3 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 neededArc Pro B70 32GB VRAM
Q419.8 GB0.08 GB1 GB20.88 GB32 GB
FP1672.0 GB0.08 GB1 GB73.08 GB32 GB

Other GPUs that run Qwen3.6-35B-A3B

All GPUs for Qwen3.6-35B-A3B →

Other models for the Arc Pro B70 32GB

All models on the Arc Pro B70 32GB →