🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN ON DATACENTER GPU
Run Spec-Drafter on Datacenter GPU hardware
13 options ranked by estimated speed at Q4 / 8k context.
| Hardware | Est tok/s | Fits 8k? | Max ctx | ~Price |
|---|---|---|---|---|
| NVIDIA B200 → | ~24,857.1 | yes | 2668k | — |
| AMD Instinct MI325X → | ~16,500 | yes | 3567k | — |
| NVIDIA H200 → | ~14,914.3 | yes | 1952k | — |
| AMD Instinct MI300X 192GB → | ~14,575 | yes | 2668k | ~$15,000 |
| NVIDIA H100 SXM5 80GB → | ~10,408.9 | yes | 1096k | ~$30,000 |
| Intel Gaudi 3 → | ~9,514.3 | yes | 1770k | — |
| NVIDIA A100 SXM4 80GB → | ~6,335.5 | yes | 1096k | ~$15,000 |
| NVIDIA H100 PCIe 80GB → | ~6,214.3 | yes | 1096k | ~$25,000 |
| NVIDIA A100 PCIe 80GB → | ~6,012.3 | yes | 1096k | ~$12,000 |
| NVIDIA RTX PRO 6000 Blackwell → | ~5,568 | yes | 1320k | ~$8,500 |
| NVIDIA RTX 6000 Ada 48GB → | ~2,982.9 | yes | 647k | ~$6,800 |
| NVIDIA L40S 48GB → | ~2,684.6 | yes | 647k | ~$10,000 |
| AMD Radeon Pro W7900 48GB → | ~2,376 | yes | 647k | ~$3,999 |
Estimate at Q4 (memory-bound). Click any option for its full per-hardware estimate and measured numbers.
FAQ
What Datacenter GPU hardware runs Spec-Drafter?▶
On GenAIList, 13 Datacenter GPU options can run Spec-Drafter at Q4 — the fastest is the NVIDIA B200 at ~24,857.1 tok/s.
How fast is Spec-Drafter on Datacenter GPU hardware?▶
Estimated generation speed ranges from ~2,376 to ~24,857.1 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.