RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can DeepSeek OCR run on the RX 7700 XT?

Yes. At 8K context the best fit is F16, using about 6.6 GB of the 11.4 GB usable on the RX 7700 XT. Estimated speed ceiling: ~215 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctxMax tok/s (est., 8K)
2-bit (Q2_K)2.0 GB Fits2.3 GB Fits1010
3-bit (Q3_K_M)2.4 GB Fits2.6 GB Fits775
4-bit (Q4_K_M) •2.7 GB Fits2.9 GB Fits648
5-bit (Q5_K_M)3.1 GB Fits3.3 GB Fits531
6-bit (Q6_K)3.4 GB Fits3.7 GB Fits462
8-bit (Q8_0) •3.8 GB Fits4.0 GB Fits404
16-bit (F16/BF16) •6.4 GB Fits6.6 GB Fits215

• real GGUF file size from sabafallah/DeepSeek-OCR-GGUF. Speed is a bandwidth ceiling (432 GB/s), not a benchmark. Methodology.

More

DeepSeek OCR on every GPU and Mac · Every model that runs on the RX 7700 XT · Other models that fit 12 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (deepseek-ai/DeepSeek-OCR), AMD specs. Data as of 2026-11-17.