← All GPUs
NVIDIA · Workstation · Blackwell
NVIDIA RTX PRO 5000 Blackwell (48GB)
Also sold as a 72 GB model (separate entry).
VRAM
48 GB
GDDR7 ECC
Memory bandwidth
1,344 GB/s
FP16 compute
not published
Price
No list price
What fits at 4-bit
The 4-bit file plus 8,192 tokens of context and 1.5 GB of overhead, against the VRAM of one card. Speeds are measured medians where someone has published a run at that card count; a blank speed means no run, not a slow one. Machine pages carry estimates for the gaps.
| Model | Needs | 1 card (48 GB) |
|---|---|---|
| Qwen3.5 4B4B · Q4_K_M | 4.3 GB | fits |
| Llama 3.1 8B Instruct8B · Q4_K_M | 7.1 GB | fits |
| Qwen3 8B8.2B · Q4_K_M | 7.3 GB | 160 tok/smedian of 1 run |
| Gemma 4 E4B IT8B · Q4_K_M | 6.5 GB | fits |
| Granite 4.2 8B8.8B · Q4_K_M | 7.9 GB | fits |
| Qwen3.5 9B9B · Q4_K_M | 7.0 GB | fits |
| Gemma 4 12B IT11.95B · Q4_K_M | 8.2 GB | fits |
| Ministral 3 14B Reasoning (2512)13.9B · Q4_K_M | 10.4 GB | fits |
| Phi-414.7B · Q4_K_M | 11.3 GB | fits |
| gpt-oss-20b21B, 3.6B active · MXFP4 (native) | 12.5 GB | 213 tok/smedian of 1 run |
| Devstral Small 2 24B (2512)24B · Q4_K_M | 16.1 GB | fits |
| Qwen3.6 27B27B · Q4_K_M | 17.7 GB | fits |
| Gemma 4 26B-A4B IT25.2B, 3.8B active · Q4_K_M | 17.4 GB | fits |
| Qwen3.8 27B27B · Q4_K_M | 18.2 GB | fits |
| Granite 4.2 30B29.3B · Q4_K_M | 20.3 GB | fits |
| Gemma 4 31B IT30.7B · Q4_K_M | 18.9 GB | fits |
| Qwen3 Coder 30B-A3B Instruct30.5B, 3.3B active · Q4_K_M | 19.5 GB | fits |
| Ornith 1.5 35B-A3B35B, 3B active · Q4_K_M | 21.9 GB | fits |
| Qwen3.6 35B-A3B35B, 3B active · Q4_K_M | 22.4 GB | fits |
| Nemotron 3.5 Lightning 30B-A3B30B, 3B active · Q4_K_M | 25.3 GB | fits |
| Llama 3.3 70B Instruct70B · Q4_K_M | 43.6 GB | 24 tok/smedian of 1 run |
| Qwen3 Coder Next80B, 3B active · Q4_K_M | 46.9 GB | fits |
Measured speeds
One card
| Model | Quant | Backend | Generation | Prompt | Context | Source |
|---|---|---|---|---|---|---|
| gpt-oss-20b | MXFP4 | llama.cpp (build not stated) | 213 tok/s | 10,340 tok/s4,096-token prompt | 4,096 | review2026-03 |
| Llama 3.3 70B Instruct | Q4_K | llama.cpp (build not stated) | 24 tok/s | 1,023 tok/s4,096-token prompt | 4,096 | review2026-03 |
| Qwen3 8B | Q4_K | llama.cpp (build not stated) | 160 tok/s | 8,175 tok/s4,096-token prompt | 4,096 | review2026-03 |
Speculative decoding, several users, or long context
Kept apart because each measures a different workload from one person chatting near the start of a conversation.
| Model | Quant | Backend | Generation | Prompt | Context | Measured on | Source |
|---|---|---|---|---|---|---|---|
| gpt-oss-20b | MXFP4 | llama.cpp (build not stated) | 196 tok/s | 8,748 tok/s16,384-token prompt | 16,384 | RTX PRO 5000 Blackwell (48GB) | review2026-03 |
| gpt-oss-20b | MXFP4 | llama.cpp (build not stated) | 183 tok/s | 7,258 tok/s32,768-token prompt | 32,768 | RTX PRO 5000 Blackwell (48GB) | review2026-03 |
| Llama 3.3 70B Instruct | Q4_K | llama.cpp (build not stated) | 22 tok/s | 826 tok/s16,384-token prompt | 16,384 | RTX PRO 5000 Blackwell (48GB) | review2026-03 |
| Qwen3 8B | Q4_K | llama.cpp (build not stated) | 126 tok/s | 5,846 tok/s16,384-token prompt | 16,384 | RTX PRO 5000 Blackwell (48GB) | review2026-03 |
| Qwen3 8B | Q4_K | llama.cpp (build not stated) | 98 tok/s | 4,006 tok/s32,768-token prompt | 32,768 | RTX PRO 5000 Blackwell (48GB) | review2026-03 |
Specifications
- Memory
- 48 GB GDDR7 ECC
- FP16 basis
- NVIDIA publishes 2,064 AI TOPS (FP4 with sparsity) and 65 TFLOPS FP32; no dense FP16 tensor figure.
- Board power
- 300 W
- Slots
- 2
- Length
- 267 mm
- PCIe
- Gen 5 x16
- NVLink
- No
- Cooling
- Blower
- Software
- CUDA, Vulkan
- Released
- 2025
- Availability
- sold new
- Price basis
- No NVIDIA list price found.
- Maker page
- nvidia.com
Sources
- nvidia.com: datasheet: 48 GB or 72 GB GDDR7 ECC 384-bit, 1344 GB/s, PCIe 5.0 x16, 300 W, 4.4 x 10.5 in dual slot, active