RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can Llama 3.1 70B run on the 2x RTX 5090?

Yes. At 8K context the best fit is Q5_K_M, using about 49.7 GB of the 60.8 GB usable on the 2x RTX 5090. Estimated speed ceiling: ~36 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •26.5 GB Fits27.7 GB Fits35.2 GB Fits65.2 GB No68
3-bit (Q3_K_M) •33.8 GB Fits35.1 GB Fits42.6 GB Fits72.6 GB No52
4-bit (Q4_K_M) •41.5 GB Fits42.7 GB Fits50.2 GB Fits80.2 GB No42
5-bit (Q5_K_M) •48.4 GB Fits49.7 GB Fits57.2 GB Tight87.2 GB No36
6-bit (Q6_K) •55.8 GB Tight57.1 GB Tight64.6 GB No94.6 GB No31
8-bit (Q8_0) •71.7 GB No73.0 GB No80.5 GB No110 GB No—
16-bit (F16/BF16) •133 GB No135 GB No142 GB No172 GB No—

• real GGUF file size from ThomasBaruzier/Llama-3.3-70B-Instruct-GGUF. Speed is a bandwidth ceiling (1792 GB/s), not a benchmark. Methodology.

More

Llama 3.1 70B on every GPU and Mac · Every model that runs on the 2x RTX 5090 · Try other settings in the calculator

Data: Hugging Face Hub (meta-llama/Llama-3.1-70B), NVIDIA specs. Data as of 2026-11-17.