Can Hunyuan A13B Instruct run on the H100 PCIe?
Yes. At 8K context the best fit is Q6_K, using about 63.0 GB of the 76.0 GB usable on the H100 PCIe. Estimated speed ceiling: ~188 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|
| 2-bit (Q2_K) • | 28.5 GB Fits | 29.0 GB Fits | 32.0 GB Fits | 420 |
| 3-bit (Q3_K_M) • | 37.1 GB Fits | 37.6 GB Fits | 40.6 GB Fits | 321 |
| 4-bit (Q4_K_M) • | 46.6 GB Fits | 47.1 GB Fits | 50.1 GB Fits | 254 |
| 5-bit (Q5_K_M) • | 54.3 GB Fits | 54.8 GB Fits | 57.8 GB Fits | 217 |
| 6-bit (Q6_K) | 62.5 GB Fits | 63.0 GB Fits | 66.0 GB Fits | 188 |
| 8-bit (Q8_0) • | 80.7 GB No | 81.2 GB No | 84.2 GB No | — |
| 16-bit (F16/BF16) • | 151 GB No | 151 GB No | 154 GB No | — |
• real GGUF file size from unsloth/Hunyuan-A13B-Instruct-GGUF. Speed is a bandwidth ceiling (2000 GB/s), not a benchmark. Methodology.
More
Hunyuan A13B Instruct on every GPU and Mac · Every model that runs on the H100 PCIe · Other models that fit 64 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (tencent/Hunyuan-A13B-Instruct), NVIDIA specs. Data as of 2026-11-17.