RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can Qwen3.5 2B run on the RTX 3080 12GB?

Yes. At 8K context the best fit is F16, using about 4.5 GB of the 11.4 GB usable on the RTX 3080 12GB. Estimated speed ceiling: ~242 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K)1.7 GB Fits1.8 GB Fits2.0 GB Fits3.2 GB Fits1069
3-bit (Q3_K_M) •2.0 GB Fits2.0 GB Fits2.3 GB Fits3.4 GB Fits824
4-bit (Q4_K_M) •2.1 GB Fits2.2 GB Fits2.4 GB Fits3.6 GB Fits712
5-bit (Q5_K_M) •2.3 GB Fits2.3 GB Fits2.6 GB Fits3.7 GB Fits635
6-bit (Q6_K) •2.4 GB Fits2.4 GB Fits2.7 GB Fits3.8 GB Fits579
8-bit (Q8_0) •2.8 GB Fits2.8 GB Fits3.1 GB Fits4.2 GB Fits453
16-bit (F16/BF16) •4.4 GB Fits4.5 GB Fits4.8 GB Fits5.9 GB Fits242

• real GGUF file size from unsloth/Qwen3.5-2B-GGUF. Speed is a bandwidth ceiling (912 GB/s), not a benchmark. Methodology.

More

Qwen3.5 2B on every GPU and Mac · Every model that runs on the RTX 3080 12GB · Other models that fit 12 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (Qwen/Qwen3.5-2B), NVIDIA specs. Data as of 2026-11-17.