Can Raon 1 nano run on the H200?
Yes. At 8K context the best fit is F16, using about 7.0 GB of the 134 GB usable on the H200. Estimated speed ceiling: ~765 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) | 2.2 GB Fits | 2.3 GB Fits | 3.0 GB Fits | 5.5 GB Fits | 3626 |
| 3-bit (Q3_K_M) | 2.6 GB Fits | 2.7 GB Fits | 3.3 GB Fits | 5.9 GB Fits | 2782 |
| 4-bit (Q4_K_M) • | 2.9 GB Fits | 3.0 GB Fits | 3.7 GB Fits | 6.2 GB Fits | 2305 |
| 5-bit (Q5_K_M) | 3.3 GB Fits | 3.4 GB Fits | 4.1 GB Fits | 6.6 GB Fits | 1909 |
| 6-bit (Q6_K) • | 3.5 GB Fits | 3.6 GB Fits | 4.2 GB Fits | 6.7 GB Fits | 1814 |
| 8-bit (Q8_0) • | 4.2 GB Fits | 4.3 GB Fits | 4.9 GB Fits | 7.4 GB Fits | 1416 |
| 16-bit (F16/BF16) • | 6.8 GB Fits | 7.0 GB Fits | 7.6 GB Fits | 10.1 GB Fits | 765 |
• real GGUF file size from KRAFTON/Raon-1-nano-gguf. Speed is a bandwidth ceiling (4800 GB/s), not a benchmark. Methodology.
More
Raon 1 nano on every GPU and Mac · Every model that runs on the H200 · Other models that fit 64 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (KRAFTON/Raon-1-nano), NVIDIA specs. Data as of 2026-11-17.