Can you run Qwen3.6 35B-A3B on RTX 3090?
Qwen3.6 35B-A3B Q4_K_M on the NVIDIA GeForce RTX 3090: verdict, VRAM math and estimated speed.
Yes, but slowly: Qwen3.6 35B-A3B spills past the RTX 3090. 22 GB weights at Q4_K_M vs 21.6 GB usable VRAM; ~53 tok/s est. (ModelFit, 2026).
VRAM math and speed are ModelFit engine estimates, not measurements. Commands are registry-verified Ollama tags.
Cite this page: ModelFit, Qwen3.6 35B-A3B on RTX 3090, https://modelfit.io/can-i-run/qwen3.6-35b-a3b-q4-on-rtx-3090/, updated September 2026, CC BY 4.0.
Last updated: September 24, 2026 · Editor: ModelFit Team
What limits this combo
The four constraints the fit engine checks for Qwen3.6 35B-A3B on the RTX 3090, in order of what usually breaks first.
Where the RTX 3090 sits for Qwen3.6 35B-A3B
The RTX 3090 is the tightest way to run Qwen3.6 35B-A3B: it loads with partial offload at ~53 tok/s est.. The cheapest card that runs it without spilling into system RAM is the RTX 5090 (32 GB, ~118 tok/s est.). Context ceiling on this card: none; on the RTX 5090: 128k.
Same engine as the verdict above: 90% of VRAM usable, KV-cache at fp16, bandwidth-derived tok/s. Verdicts of the other cards are computed for Qwen3.6 35B-A3B Q4_K_M exactly.
Quant explorer
Switch between the quality-gated builds tracked for this exact combo. Q4_K_M stays the default recommendation; heavier Q6/Q8 builds only appear when their exact Ollama tags are registry-verified.
Why no 2-bit builds: Q1/Q2-class quants can look attractive in a memory table, but their quality loss is large enough that ModelFit excludes them from rankings instead of inflating the catalog with junk options. Usable VRAM: 21.6 GB.
Workload verdicts
Qwen3.6 35B-A3B on the RTX 3090, graded per use case from the model's registry-verified tuning and the engine's speed estimate for this exact combo.
Grades combine tag-verified tuning (a model not built for the workload caps at C) with the tok/s floor each workload needs to feel usable. A combo that partially offloads caps at C; one that does not fit is D everywhere.
Memory math: weights + context vs budget
Weights take 22 GB. Context costs extra KV-cache on top; this is where long-context sessions break on cards that technically fit the weights.
This configuration exceeds the comfortable VRAM budget by about 0.7 GB. Expect offload pressure or a shorter safe context.
KV-cache figures assume an fp16 cache, the llama.cpp/Ollama default. Standard GQA models use a size-class estimate (8 KV heads x 128 head dim class); hybrid linear-attention models (Qwen3.5/3.6, Qwen3-Next) use the exact per-token cost from their published config, since only their sparse full-attention layers cache KV. A q8_0 KV cache roughly halves either figure. Estimates, not measurements.
See how fast it feels
A deterministic typing simulation for Qwen3.6 35B-A3B: first token ~1.0s (instant prefill), then ~53 tokens/sec.
ModelFit engine estimate, not a measured benchmark. Real speed varies with prompt length, thermals, runtime, and KV-cache settings.
Upgrade path
The cheapest tracked card that runs Qwen3.6 35B-A3B comfortably is the AMD Ryzen AI Max+ 395 (Strix Halo) (110 GB unified memory).
See the Ryzen AI Max+ 395 pageQwen3.6 35B-A3B on RTX 3090: FAQ
Can the NVIDIA GeForce RTX 3090 run Qwen3.6 35B-A3B?
Barely. Qwen3.6 35B-A3B (Q4_K_M) needs about 22 GB but the RTX 3090 has 21.6 GB usable VRAM, so part of the model spills to system RAM and speed drops to at roughly 53 tokens/sec (est.).
How much VRAM does Qwen3.6 35B-A3B need?
About 22 GB for the weights at Q4_K_M, plus KV-cache for context: roughly 0.3 GB extra at 16k tokens. The RTX 3090 budget is 21.6 GB (24 GB x 90%).
What is the best quantization of Qwen3.6 35B-A3B for the RTX 3090?
Stick with the Q4_K_M build at 22 GB; every heavier quant exceeds the 21.6 GB usable VRAM.
What GPU do I need to run Qwen3.6 35B-A3B comfortably?
The cheapest tracked card that runs Qwen3.6 35B-A3B (Q4_K_M) fully in VRAM is the AMD Ryzen AI Max+ 395 (Strix Halo) (110 GB).
How much context can Qwen3.6 35B-A3B use on the RTX 3090?
None worth having: the 22 GB of weights alone exceed the 21.6 GB usable VRAM, so the model only runs with system-RAM offload and short prompts.
Is Qwen3.6 35B-A3B on the RTX 3090 fast enough for coding agents?
Yes: the engine estimates ~53 tok/s against the ~30 tok/s an agentic loop needs to feel responsive, and Qwen3.6 35B-A3B is tuned for tool-calling.