RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can deepseek coder 1.3b instruct run on the 2x RTX 4090?

Yes. At 8K context the best fit is F16, using about 4.5 GB of the 45.6 GB usable on the 2x RTX 4090. Estimated speed ceiling: ~374 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •1.8 GB Fits2.5 GB Fits1596
3-bit (Q3_K_M) •1.9 GB Fits2.6 GB Fits1430
4-bit (Q4_K_M) •2.0 GB Fits2.8 GB Fits1154
5-bit (Q5_K_M) •2.1 GB Fits2.9 GB Fits1006
6-bit (Q6_K) •2.3 GB Fits3.1 GB Fits860
8-bit (Q8_0) •2.5 GB Fits3.3 GB Fits704
16-bit (F16/BF16)3.7 GB Fits4.5 GB Fits374

• real GGUF file size from TheBloke/deepseek-coder-1.3b-instruct-GGUF. Speed is a bandwidth ceiling (1008 GB/s), not a benchmark. Methodology.

More

deepseek coder 1.3b instruct on every GPU and Mac · Every model that runs on the 2x RTX 4090 · Try other settings in the calculator

Data: Hugging Face Hub (deepseek-ai/deepseek-coder-1.3b-instruct), NVIDIA specs. Data as of 2026-11-17.