🔌
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 Phi-3.5-MoE 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 → | ~204.4 | yes | 431k | — |
| AMD Instinct MI325X → | ~135.7 | yes | 610k | — |
| NVIDIA H200 → | ~122.7 | yes | 287k | — |
| AMD Instinct MI300X 192GB → | ~119.9 | yes | 431k | ~$15,000 |
| NVIDIA H100 SXM5 80GB → | ~85.6 | yes | 116k | ~$30,000 |
| Intel Gaudi 3 → | ~78.2 | yes | 251k | — |
| NVIDIA A100 SXM4 80GB → | ~52.1 | yes | 116k | ~$15,000 |
| NVIDIA H100 PCIe 80GB → | ~51.1 | yes | 116k | ~$25,000 |
| NVIDIA A100 PCIe 80GB → | ~49.4 | yes | 116k | ~$12,000 |
| NVIDIA RTX PRO 6000 Blackwell → | ~45.8 | yes | 161k | ~$8,500 |
| NVIDIA RTX 6000 Ada 48GB → | ~24.5 | yes | 26k | ~$6,800 |
| NVIDIA L40S 48GB → | ~22.1 | yes | 26k | ~$10,000 |
| AMD Radeon Pro W7900 48GB → | ~19.5 | yes | 26k | ~$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 Phi-3.5-MoE?▶
On GenAIList, 13 Datacenter GPU options can run Phi-3.5-MoE at Q4 — the fastest is the NVIDIA B200 at ~204.4 tok/s.
How fast is Phi-3.5-MoE on Datacenter GPU hardware?▶
Estimated generation speed ranges from ~19.5 to ~204.4 tok/s at Q4 (memory-bandwidth bound). Click any option for its full estimate and any measured numbers.