NVIDIA Tesla P40 vs NVIDIA GeForce RTX 5090
For running models locally, priced in the US. At 64k context the first runs 22 models fully in memory, the second 27.
Specifications
| Spec | NVIDIA Tesla P40 | NVIDIA GeForce RTX 5090 |
|---|---|---|
| GPU | Tesla P40 | GeForce RTX 5090 |
| Memory | 24 GB GDDR5 | 32 GB GDDR7 |
| Memory bus | 384-bit | 512-bit |
| Memory bandwidth | 346 GB/s | 1,792 GB/s |
| FP16 compute | 11.8 TFLOPS | 209 TFLOPS |
| Software | CUDA | CUDA |
| Board power | 250 W | 575 W |
| PCIe generation | Gen 3 | Gen 5 |
| PCIe lanes | x16 | x16 |
What each runs
Estimated decode speed with the most precise quant that fits each one est.
| Model | NVIDIA Tesla P40 | NVIDIA GeForce RTX 5090 |
|---|---|---|
| Qwen 3.8 27B | 10tok/sUD-Q4_K_M | 42tok/sUD-Q6_K |
| Qwen 3.6 27B | 9tok/sQ5_K_M | 41tok/sQ6_K |
| Gemma 4 31B | offload | 35tok/sQ6_K |
| Qwen 3.6 35B-A3B | 40tok/sUD-Q4_K_M | 173tok/sUD-Q6_K |
| Qwen 3.5 27B | 10tok/sQ4_K_M | 41tok/sQ6_K |
| Gemma 4 26B-A4B | 27tok/sUD-Q5_K_M | 117tok/sQ8_0 |
| Qwen 3.5 35B-A3B | 40tok/sQ4_K_M | 175tok/sQ6_K |
| GLM 4.7 Flash | 20tok/sQ4_K_M | 87tok/sQ6_K |
| Devstral Small 2 24B | offload | 37tok/sQ6_K |
| gpt-oss 20B | 32tok/sMXFP4 | 165tok/sMXFP4 |
| Kimi Linear 48B-A3B | offload | 277tok/sQ4_K_M |
| Qwen3 Coder 30B-A3B | offload | 75tok/sQ6_K |
| Gemma 4 12B | 15tok/sQ8_0 | 79tok/sQ8_0 |
| Qwen 3.5 9B | 18tok/sQ8_0 | 95tok/sQ8_0 |
| Gemma 3 27B | 10tok/sQ4_K_M | 40tok/sQ6_K |
| Mistral Small 3.2 24B | offload | 37tok/sQ6_K |
| Ministral 3 14B | 10tok/sQ6_K | 44tok/sQ8_0 |
| Qwen3 14B | 9tok/sQ6_K | 42tok/sQ8_0 |
| DeepSeek R1 0528 Qwen3 8B | 12tok/sQ8_0 | 61tok/sQ8_0 |
| Qwen3 8B | 12tok/sQ8_0 | 61tok/sQ8_0 |
| Phi-4 14B | 9tok/sQ5_K_M | 38tok/sQ8_0 |
| Qwen 3.5 4B | 32tok/sQ8_0 | 166tok/sQ8_0 |
| Ministral 3 8B | 12tok/sQ8_0 | 61tok/sQ8_0 |
| Granite 4.2 8B | 11tok/sQ8_0 | 55tok/sQ8_0 |
| Gemma 4 E4B | 21tok/sQ8_0 | 110tok/sQ8_0 |
| LFM2.5 8B-A1B | 30tok/sQ8_0 | 156tok/sQ8_0 |
| Llama 3.1 8B | 14tok/sQ6_K | 73tok/sQ6_K |
| Qwen 3.5 122B-A10B | offload | offload |
| Qwen3 Coder Next | offload | offload |
| Nemotron 3 Super 120B-A12B | won't run | offload |
| Mistral Small 4 119B | offload | offload |
| GLM 4.5 Air | won't run | offload |
| Qwen3 Next 80B-A3B | offload | offload |
| Nemotron 3.5 Lightning 30B-A3B | offload | offload |
| Qwen3 32B | offload | offload |
| DeepSeek R1 Distill Qwen 32B | offload | offload |
| DeepSeek V4 Flash | won't run | won't run |
| GLM 5.3 Flash | won't run | won't run |
| Qwen 3.8 Flash Next | won't run | won't run |
| MiniMax M2.7 | won't run | won't run |
| gpt-oss 120B | won't run | won't run |
| Llama 3.3 70B | won't run | won't run |