RTX 5090 vs Spark x4
Best verified output tok/s per model (single-stream generation). Hover a value for the quant.
RTX 509024 verified runs · best 2231.4 tok/s
Spark x41 verified run · best 30.2 tok/s
| Model | RTX 5090 | Spark x4 |
|---|---|---|
| LFM2-350M-NVFP4A16 | 2231.4 | — |
| NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 | 446.7 | — |
| Qwen3.8-27B | 435.1 | — |
| Ornith-1.0-35B-PrismaAURA-4.75bit-vllm-MTP | 283.1 | — |
| Kimi-Linear-48B-A3B-Instruct | 280.8 | — |
| Ornith-1.5-35B-A3B-NVFP4 | 258.8 | — |
| North-Mini-Code-1.0 | 258.4 | — |
| Qwen3.6-35B-A3B-MTP-GGUF | 222.9 | — |
| Laguna-XS-2.1-NVFP4 | 216.7 | — |
| Qwen3.8-27B-nvfp4-NInfer | 213.0 | — |
| Laguna-XS-2.1-INT4 | 206.2 | — |
| KAT-Coder-V2.5-Dev-AWQ-W4A16-ASYM | 201.2 | — |
| Ornith-1.0-35B-AWQ-INT4 | 196.6 | — |
| Qwen3.8-27B-GPTQ-W4A16 | 137.9 | — |
| Muse-Glimmer-30B-NVFP4 | 124.8 | — |
| Qwen3.6-27B-NVFP4 | 120.9 | — |
| Huihui-Qwen3.8-27B-abliterated-NVFP4 | 117.7 | — |
| Qwen3.8-27B-W4A16-AWQ | 108.0 | — |
| Qwen3.8-27B-5090-goldilocks-GGUF | 93.5 | — |
| Qwen3.8-27B-GGUF | 79.4 | — |
| DeepSeek-V4.1-Flash | — | 30.2 |
| Qwen3.6-27B-MTP-GGUF | 28.3 | — |
| DeepSeek-V4-Flash-0731-MXFP4-GGUF | 26.4 | — |
| Qwen3.8-27B-FP8 | 16.8 | — |
| Qwen3.6-27B-GGUF | 4.1 | — |