QwenQwen3.6-35B-A3B on NVIDIA Grace Blackwell Ultra GB300
Yes — Grace Blackwell Ultra GB300 runs Qwen3.6-35B-A3B excellently at Q5_K_M — 145 tok/s. 288 GB VRAM with plenty of headroom.
Model Size
35B
Device VRAM
288 GB
Bandwidth
8000 GB/s
Quantization
Q5_K_M
Performance by Quantization
OwnRig currently has one published compatibility entry for Qwen3.6-35B-A3B on NVIDIA Grace Blackwell Ultra GB300 at Q5_K_M. This is the best supported pairing we can stand behind today.
| Quantization | Speed | TTFT | Fits in VRAM | Rating | Confidence |
|---|---|---|---|---|---|
| Q5_K_M | 145 tok/s | 70ms | ✓ Yes | Excellent | estimated |
Notes
Q5_K_M
Qwen 3.6-35B-A3B MoE (~3B active). Q5_K_M on 288GB gb300-288gb.
About Qwen3.6-35B-A3B
Qwen3.6-35B-A3B (35B) is a chat, coding, ai coding, reasoning, multi-purpose model. MoE with 35B total parameters and ~3B active per token. Local stacks still load the full expert pool, so VRAM tracks total size. Q4_K_M needs ~22GB at practical context (RTX 4090 / 24GB class sweet spot). Q3_K_M (~16GB) fits 16GB cards at default context with minimal KV headroom; long 262K runs need KV cache quantization. 8GB cards cannot run this model without heavy CPU offload. Apache 2.0.
View all Qwen3.6-35B-A3B 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 Unsloth GGUF sizes; best quant that fits device VRAM. Reference hardware source: huggingface.co (2026-05-26)
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.