RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can Phi 4 mini instruct run on the L40S?

Yes. At 8K context the best fit is F16, using about 8.9 GB of the 45.6 GB usable on the L40S. Estimated speed ceiling: ~112 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •2.8 GB Fits3.3 GB Fits6.3 GB Fits18.3 GB Fits513
3-bit (Q3_K_M) •3.3 GB Fits3.8 GB Fits6.8 GB Fits18.8 GB Fits408
4-bit (Q4_K_M) •3.6 GB Fits4.1 GB Fits7.1 GB Fits19.1 GB Fits347
5-bit (Q5_K_M) •3.9 GB Fits4.4 GB Fits7.4 GB Fits19.4 GB Fits303
6-bit (Q6_K) •4.2 GB Fits4.7 GB Fits7.7 GB Fits19.7 GB Fits274
8-bit (Q8_0) •5.1 GB Fits5.6 GB Fits8.6 GB Fits20.6 GB Fits212
16-bit (F16/BF16) •8.4 GB Fits8.9 GB Fits11.9 GB Fits23.9 GB Fits112

• real GGUF file size from unsloth/Phi-4-mini-instruct-GGUF. Speed is a bandwidth ceiling (864 GB/s), not a benchmark. Methodology.

More

Phi 4 mini instruct on every GPU and Mac · Every model that runs on the L40S · Other models that fit 48 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (microsoft/Phi-4-mini-instruct), NVIDIA specs. Data as of 2026-11-17.