Can Nemotron 3.5 Parse run on the RX 9070?
Yes. At 8K context the best fit is F16, using about 1.9 GB of the 15.2 GB usable on the RX 9070. Estimated speed ceiling: ~490 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | Max tok/s (est., 8K) |
|---|---|---|---|
| 2-bit (Q2_K) | 0.8 GB Fits | 0.9 GB Fits | 2706 |
| 3-bit (Q3_K_M) | 0.9 GB Fits | 1.0 GB Fits | 2076 |
| 4-bit (Q4_K_M) | 1.0 GB Fits | 1.1 GB Fits | 1660 |
| 5-bit (Q5_K_M) | 1.0 GB Fits | 1.1 GB Fits | 1424 |
| 6-bit (Q6_K) | 1.1 GB Fits | 1.2 GB Fits | 1237 |
| 8-bit (Q8_0) • | 1.4 GB Fits | 1.5 GB Fits | 739 |
| 16-bit (F16/BF16) • | 1.8 GB Fits | 1.9 GB Fits | 490 |
• real GGUF file size from SkyContact/Nemotron-3.5-Parse-GGUF. Speed is a bandwidth ceiling (640 GB/s), not a benchmark. Methodology.
More
Nemotron 3.5 Parse on every GPU and Mac · Every model that runs on the RX 9070 · Other models that fit 16 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (nvidia/Nemotron-3.5-Parse), AMD specs. Data as of 2026-11-17.