Certified Models by Hardware
Spectro Cloud certifies a small set of large language models (LLMs) for the hardware that PaletteAI Inference Launchpad supports. A certified model is one that Spectro Cloud has validated to run correctly on the listed GPU configuration. For what certification covers, refer to Model Certification.
info
This is not an exclusive list. You can load other models on the appliance as long as they fit within the available GPU resources. If the model you want does not appear here, contact Spectro Cloud to discuss your use case.
- NVIDIA
- AMD
| GPU configuration | GLM 5.2 | DeepSeek v4 Pro | Kimi 2.7 | Gemma 4 |
|---|---|---|---|---|
| 4 x H100 | ❌ | ❌ | ❌ | ✅ |
| 8 x H100 | ❌ | ❌ | ❌ | ✅ |
| 4 x H200 | ❌ | ❌ | ❌ | ✅ |
| 8 x H200 | ✅ | ✅ | ✅ | ✅ |
| 4 x B200 | ❌ | ❌ | ❌ | ✅ |
| 8 x B200 | ✅ | ✅ | ✅ | ✅ |
| GPU configuration | GLM 5.2 | DeepSeek v4 Pro | Kimi 2.7 | Gemma 4 |
|---|---|---|---|---|
| 4 x MI300X | ❌ | ❌ | ✅ | ✅ |
| 8 x MI300X | ✅ | ❌ | ✅ | ✅ |
| 4 x MI325X | ✅ | ❌ | ✅ | ✅ |
| 8 x MI325X | ✅ | ❌ | ❌ | ✅ |
| 4 x MI350X | ✅ | ✅ | ❌ | ✅ |
| 8 x MI350X | ✅ | ✅ | ❌ | ✅ |
| 4 x MI355X | ✅ | ✅ | ❌ | ✅ |
| 8 x MI355X | ✅ | ✅ | ❌ | ✅ |