RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can Qwen2.5 Coder 7B Instruct run on the RTX 5090?

Yes. At 8K context the best fit is F16, using about 15.3 GB of the 30.4 GB usable on the RTX 5090. Estimated speed ceiling: ~118 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctxMax tok/s (est., 8K)
2-bit (Q2_K)3.6 GB Fits3.8 GB Fits5.1 GB Fits627
3-bit (Q3_K_M)4.4 GB Fits4.6 GB Fits5.9 GB Fits481
4-bit (Q4_K_M)5.2 GB Fits5.5 GB Fits6.8 GB Fits385
5-bit (Q5_K_M)6.0 GB Fits6.2 GB Fits7.5 GB Fits330
6-bit (Q6_K)6.7 GB Fits6.9 GB Fits8.3 GB Fits287
8-bit (Q8_0)8.4 GB Fits8.7 GB Fits10.0 GB Fits221
16-bit (F16/BF16)15.1 GB Fits15.3 GB Fits16.6 GB Fits118

• real GGUF file size. Speed is a bandwidth ceiling (1792 GB/s), not a benchmark. Methodology.

More

Qwen2.5 Coder 7B Instruct 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/Qwen2.5-Coder-7B-Instruct), NVIDIA specs. Data as of 2026-11-17.