RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can Qwen3.8 27B run on the RTX 5090?

Yes. At 8K context the best fit is Q6_K, using about 21.8 GB of the 30.4 GB usable on the RTX 5090. Estimated speed ceiling: ~82 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K)10.8 GB Fits11.1 GB Fits12.6 GB Fits18.6 GB Fits172
3-bit (Q3_K_M)13.8 GB Fits14.0 GB Fits15.5 GB Fits21.5 GB Fits132
4-bit (Q4_K_M) •16.6 GB Fits16.8 GB Fits18.3 GB Fits24.3 GB Fits108
5-bit (Q5_K_M) •19.6 GB Fits19.9 GB Fits21.4 GB Fits27.4 GB Tight90
6-bit (Q6_K) •21.6 GB Fits21.8 GB Fits23.3 GB Fits29.3 GB Tight82
8-bit (Q8_0) •28.2 GB Tight28.4 GB Tight29.9 GB Tight35.9 GB No62
16-bit (F16/BF16) •52.0 GB No52.3 GB No53.8 GB No59.8 GB No—

• real GGUF file size from unsloth/Qwen3.8-27B-GGUF. Speed is a bandwidth ceiling (1792 GB/s), not a benchmark. Methodology.

More

Qwen3.8 27B on every GPU and Mac · Every model that runs on the RTX 5090 · Other models that fit 32 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (Qwen/Qwen3.8-27B), NVIDIA specs. Data as of 2026-11-17.