QwenQwen3.5-397B (MoE) on NVIDIA Grace Blackwell Ultra GB300
Yes — Grace Blackwell Ultra GB300 runs Qwen3.5-397B (MoE) excellently at Q4_K_M — 120 tok/s. 288 GB VRAM with plenty of headroom.
Model Size
397B
Device VRAM
288 GB
Bandwidth
8000 GB/s
Quantization
Q4_K_M
Performance by Quantization
OwnRig currently has one published compatibility entry for Qwen3.5-397B (MoE) on NVIDIA Grace Blackwell Ultra GB300 at Q4_K_M. This is the best supported pairing we can stand behind today.
| Quantization | Speed | TTFT | Fits in VRAM | Rating | Confidence |
|---|---|---|---|---|---|
| Q4_K_M | 120 tok/s | 67ms | ✓ Yes | Excellent | estimated |
Notes
Q4_K_M
Q4_K_M fits in 288GB with 58GB headroom. Frontier MoE model: 397B total params, ~30B active per token. Consumer hardware requires heavy offloading.
About Qwen3.5-397B (MoE)
Qwen3.5-397B (MoE) (397B) is a chat, coding, ai coding, reasoning, multi-purpose model. Frontier-scale MoE with 397B total parameters and roughly 30B active per token. Requires an enormous memory footprint even when quantized, so local runs are limited to GB300-class hardware or aggressive offloading setups.
View all Qwen3.5-397B (MoE) hardware options →About NVIDIA Grace Blackwell Ultra GB300
NVIDIA Grace Blackwell Ultra GB300 has 288 GB at 8000 GB/s. Street price: $30,000.
See all models NVIDIA Grace Blackwell Ultra GB300 can run →Estimate method: Estimated from frontier MoE reports and bandwidth scaling for ~30B-active models. Reference hardware source: github.com (2026-03-24)
Performance varies by driver version, inference engine, quantization method, context length, and system configuration. Figures shown are estimates based on community benchmarks and may not reflect your exact setup. Product names are trademarks of their respective owners. OwnRig is independent and not affiliated with any hardware or AI model provider.