RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can Olmo 3 7B Instruct run on the L40S?

Yes. At 8K context the best fit is F16, using about 16.7 GB of the 45.6 GB usable on the L40S. Estimated speed ceiling: ~59 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •5.3 GB Fits5.8 GB Fits8.8 GB Fits302
3-bit (Q3_K_M) •6.0 GB Fits6.5 GB Fits9.5 GB Fits237
4-bit (Q4_K_M) •6.8 GB Fits7.3 GB Fits10.3 GB Fits193
5-bit (Q5_K_M) •7.4 GB Fits7.9 GB Fits10.9 GB Fits166
6-bit (Q6_K) •8.2 GB Fits8.7 GB Fits11.7 GB Fits144
8-bit (Q8_0) •9.8 GB Fits10.3 GB Fits13.3 GB Fits111
16-bit (F16/BF16) •16.2 GB Fits16.7 GB Fits19.7 GB Fits59

• real GGUF file size from unsloth/Olmo-3-7B-Instruct-GGUF. Speed is a bandwidth ceiling (864 GB/s), not a benchmark. Methodology.

More

Olmo 3 7B 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 (allenai/Olmo-3-7B-Instruct), NVIDIA specs. Data as of 2026-11-17.