RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run Qwen3.6 35B A3B locally?

Qwen3.6 35B A3B by Qwen has 36B parameters (3.0B active per token, 256 experts, 8 used per token). At 4-bit with an 8K context it needs about 22.1 GB of GPU memory. It fits on 48 of 65 GPUs and Macs we track. See all models that fit in 24 GB of VRAM.

22.1 GBQ4_K_M · 8K context
RTX 5090smallest consumer GPU that fits
M4 Pro 48GBsmallest Mac that fits
256Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/Qwen3.6-35B-A3B-MTP-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K)rough13.5 GB13.6 GB14.1 GB15.9 GB
3-bit (Q3_K_M) •usable16.9 GB17.0 GB17.4 GB19.3 GB
4-bit (Q4_K_M) •good, most popular22.1 GB22.1 GB22.6 GB24.5 GB
5-bit (Q5_K_M) •very good26.2 GB26.3 GB26.7 GB28.6 GB
6-bit (Q6_K) •near lossless28.9 GB29.0 GB29.4 GB31.3 GB
8-bit (Q8_0) •effectively lossless36.2 GB36.2 GB36.7 GB38.6 GB
16-bit (F16/BF16) •original weights67.1 GB67.2 GB67.7 GB69.6 GB

Which GPUs and Macs run Qwen3.6 35B A3B?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBNo———details
RTX 4060 Ti 16GB16 GBFitsQ2_K13.6 GB256details
RTX 309024 GBFitsQ3_K_M17.0 GB656details
RTX 409024 GBFitsQ3_K_M17.0 GB706details
RTX 509032 GBFitsQ5_K_M26.3 GB793details
RTX 5070 Ti16 GBFitsQ2_K13.6 GB796details
RX 7900 XTX24 GBFitsQ3_K_M17.0 GB673details
Mac M4 Pro 48GB48 GBFitsQ5_K_M26.3 GB121details
Mac M4 Max 128GB128 GBFitsF1667.2 GB92details
Mac M3 Ultra 512GB512 GBFitsF1667.2 GB138details
DGX Spark128 GBFitsF1667.2 GB46details
RTX 407012 GBNo———details
RTX 5060 Ti 16GB16 GBFitsQ2_K13.6 GB398details
RTX 40608 GBNo———details
Arc B58012 GBNo———details
RTX 3060 Ti8 GBNo———details
RTX 30708 GBNo———details
RTX 3080 10GB10 GBNo———details
RTX 3080 12GB12 GBNo———details
RTX 3090 Ti24 GBFitsQ3_K_M17.0 GB706details
RTX 4060 Ti 8GB8 GBNo———details
RTX 4070 Super12 GBNo———details
RTX 4070 Ti12 GBNo———details
RTX 4070 Ti Super16 GBFitsQ2_K13.6 GB597details
RTX 408016 GBFitsQ2_K13.6 GB637details
RTX 4080 Super16 GBFitsQ2_K13.6 GB654details
RTX 50608 GBNo———details
RTX 507012 GBNo———details
RTX 508016 GBFitsQ2_K13.6 GB853details
2x RTX 30902×24 GBFitsQ8_036.2 GB297details
2x RTX 40902×24 GBFitsQ8_036.2 GB320details
2x RTX 50902×32 GBFitsQ8_036.2 GB568details
RTX A600048 GBFitsQ8_036.2 GB243details
RTX 6000 Ada48 GBFitsQ8_036.2 GB304details
RTX PRO 6000 Blackwell96 GBFitsF1667.2 GB302details
L40S48 GBFitsQ8_036.2 GB274details
A100 40GB40 GBFitsQ6_K29.0 GB621details
A100 80GB80 GBFitsF1667.2 GB344details
H100 PCIe80 GBFitsF1667.2 GB337details
H100 SXM80 GBFitsF1667.2 GB565details
H200141 GBFitsF1667.2 GB809details
RX 76008 GBNo———details
RX 7700 XT12 GBNo———details
RX 7800 XT16 GBFitsQ2_K13.6 GB555details
RX 7900 GRE16 GBFitsQ2_K13.6 GB512details
RX 7900 XT20 GBFitsQ3_K_M17.0 GB561details
RX 907016 GBFitsQ2_K13.6 GB569details
RX 9070 XT16 GBFitsQ2_K13.6 GB569details
Ryzen AI Max+ 395 128GB128 GBFitsF1667.2 GB43details
Arc A770 16GB16 GBFitsQ2_K13.6 GB498details
Mac M1 16GB16 GBNo———details
Mac M2 24GB24 GBFitsQ2_K13.6 GB89details
Mac M3 24GB24 GBFitsQ2_K13.6 GB89details
Mac M4 16GB16 GBNo———details
Mac M4 32GB32 GBFitsQ3_K_M17.0 GB84details
Mac M1 Pro 32GB32 GBFitsQ3_K_M17.0 GB140details
Mac M3 Pro 36GB36 GBFitsQ3_K_M17.0 GB105details
Mac M4 Pro 24GB24 GBFitsQ2_K13.6 GB243details
Mac M4 Pro 64GB64 GBFitsQ8_036.2 GB87details
Mac M1 Max 64GB64 GBFitsQ8_036.2 GB127details
Mac M2 Max 96GB96 GBFitsQ8_036.2 GB127details
Mac M3 Max 128GB128 GBFitsF1667.2 GB67details
Mac M4 Max 64GB64 GBFitsQ8_036.2 GB173details
Mac M2 Ultra 192GB192 GBFitsF1667.2 GB135details
Mac M3 Ultra 256GB256 GBFitsF1667.2 GB138details

Prefer an API?

Qwen: Qwen3.6 35B A3B is on OpenRouter at $0.10 per 1M input tokens and $0.90 per 1M output tokens. Price details and history →

Questions

How much VRAM does Qwen3.6 35B A3B need?

About 22.1 GB at 4-bit (Q4_K_M) with an 8K context: 21.1 GB of weights (real GGUF file size), 0.2 GB of KV cache and 0.9 GB of runtime overhead. At 8-bit it needs about 36.2 GB.

What is the smallest GPU that runs Qwen3.6 35B A3B?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 5090 (32 GB).

Can I run Qwen3.6 35B A3B on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M4 Pro 48GB or larger, using the default macOS GPU memory limit.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

Qwen3.5 35B A3B

36B · 21.5 GB at 4-bit

Qwen AgentWorld 35B A3B

35B · 21.6 GB at 4-bit

Qwen3 32B

33B · 21.1 GB at 4-bit

Ornith 1.5 35B A3B

36B · 21.4 GB at 4-bit

Hermes 4.3 36B

36B · 23.0 GB at 4-bit

Seed OSS 36B Instruct

36B · 23.0 GB at 4-bit

Model facts

Hugging Face repoQwen/Qwen3.6-35B-A3B
Parameters35,951,822,704 (36B) · ~3.0B active (from the model name)
Layers / hidden size40 / 2048
AttentionGQA (hybrid: 30 linear/state layers); 16 heads, 2 KV heads, head dim 256
KV cache per token (FP16)20.0 KiB
Max context262,144 tokens
Licenseapache-2.0
Released on HFApr 15, 2026
Downloads (30d) / likes3,908,022 / 3,137

Sources: Hugging Face model API · config.json (Qwen/Qwen3.6-35B-A3B) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.