RTX PRO 5000 72GB Blackwell
NVIDIA RTX PRO 5000 72GB Blackwell: 72 GB of memory at 1344 GB/s. Below: the open models it can run, estimated at 72 GB with an 8K context. 72 GB of VRAM at 1,344 GB/s in a 300 W card: room for 100B-class MoE models such as gpt-oss-120b at 4-bit. NVIDIA has not published a price.
Last updated: 2026-10-14
Specifications
What it can run
Best-quality quantization that fits with a 8K context, and the estimated generation speed. Estimates, not benchmarks. See the methodology.
| Model | Params | Best quant | Memory | tok/s (est.) |
|---|---|---|---|---|
| Mistral Small 4 119B (2603) | 119B (6.5B) | Q4_K_M | 71.2 GiB | 239 |
| gpt-oss-120b | 117B (5.1B) | Q4_K_M | 70.1 GiB | 304 |
| Llama 4 Scout 17B-16E Instruct | 109B (17B) | Q4_K_M | 66.6 GiB | 91 |
| GLM-4.5-Air | 106B (12B) | Q4_K_M | 64.8 GiB | 129 |
| Qwen3-Coder-Next (80B-A3B) | 80B (3B) | Q6_K | 65.1 GiB | 381 |
| Llama 3.3 70B Instruct | 70B | Q6_K | 59.4 GiB | 16 |
| Qwen3.5-35B-A3B | 35B (3B) | FP16 | 69.1 GiB | 157 |
| Qwen3.6-35B-A3B | 35B (3B) | FP16 | 69.1 GiB | 157 |
| Qwen3-32B | 32.8B | FP16 | 66.6 GiB | 14 |
| DeepSeek-R1-Distill-Qwen-32B | 32.76B | FP16 | 66.6 GiB | 14 |
| Gemma 4 31B IT | 30.7B | FP16 | 61.6 GiB | 15 |
| Qwen3-30B-A3B (2507) | 30.5B (3.3B) | FP16 | 60.9 GiB | 143 |
| Qwen3-Coder-30B-A3B-Instruct | 30.5B (3.3B) | FP16 | 60.9 GiB | 143 |
| GLM-4.7-Flash (30B-A3B) | 30B (3B) | FP16 | 59.6 GiB | 157 |
| Qwen3.5-27B | 27B | FP16 | 53.8 GiB | 17 |
| Qwen3.6-27B | 27B | FP16 | 53.8 GiB | 17 |
| Qwen3.8-27B | 27B | FP16 | 53.8 GiB | 17 |
| Gemma 3 27B IT | 27B | FP16 | 54.3 GiB | 17 |
| Gemma 4 26B A4B IT | 25.2B (3.8B) | FP16 | 50.1 GiB | 124 |
| Mistral Small 3.2 24B Instruct (2506) | 24B | FP16 | 48.7 GiB | 20 |
| Magistral Small 1.2 (2509) | 24B | FP16 | 48.7 GiB | 20 |
| Devstral Small 2 24B Instruct (2512) | 24B | FP16 | 48.7 GiB | 20 |
| gpt-oss-20b | 21B (3.6B) | FP16 | 41.8 GiB | 131 |
| Phi-4 (14B) | 14B | FP16 | 29.4 GiB | 34 |
| Gemma 3 12B IT | 12B | FP16 | 24.8 GiB | 39 |
| Gemma 4 12B Unified IT | 11.95B | FP16 | 24.2 GiB | 39 |
| Qwen3.5-9B | 9B | FP16 | 18.4 GiB | 52 |
| Qwen3-8B | 8.2B | FP16 | 17.7 GiB | 57 |
| DeepSeek-R1-0528-Qwen3-8B | 8.19B | FP16 | 17.6 GiB | 57 |
| Llama 3.1 8B Instruct | 8B | FP16 | 17.1 GiB | 59 |
| Gemma 4 E4B IT | 8B | FP16 | 16.4 GiB | 59 |
| Llama 3.2 3B Instruct | 3.21B | FP16 | 7.7 GiB | 147 |
| Kimi K2.6 | 1000B (32B) | Doesn't fit | – | – |
| DeepSeek-R1-0528 (671B) | 671B (37B) | Doesn't fit | – | – |
| DeepSeek-V4-Flash | 284B (13B) | Doesn't fit | – | – |
| Qwen3.8-Flash-Next | 180B (6B) | Doesn't fit | – | – |
| Qwen3.5-122B-A10B | 122B (10B) | Doesn't fit | – | – |
Prices
No prices recorded for this device yet. All prices →
Frequently asked questions
Which LLMs can the RTX PRO 5000 72GB Blackwell run?
With 72 GB and an 8K context, the largest models that fit include: Mistral Small 4 119B (2603), gpt-oss-120b, Llama 4 Scout 17B-16E Instruct, GLM-4.5-Air, Qwen3-Coder-Next (80B-A3B). See the full table below.
How fast is the RTX PRO 5000 72GB Blackwell for local LLMs?
Generation speed is mostly set by memory bandwidth (1344 GB/s). For Llama 3.1 8B Instruct at FP16, expect roughly 59 tokens/s; bigger models are slower.
Can the RTX PRO 5000 72GB Blackwell run 70B models?
Yes. Llama 3.3 70B Instruct fits on the RTX PRO 5000 72GB Blackwell (72 GB) at Q6_K with an 8K context, using about 59.4 GiB, at an estimated 16 tokens/s.
Sources
- https://www.nvidia.com/content/dam/en-zz/Solutions/products/workstations/professional-desktop-gpus/rtx-pro-5000-blackwell/workstation-datasheet-blackwell-rtx-pro-5000-5488550-nvidia.pdf
- https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/rtx-pro-5000/
- https://blogs.nvidia.com/blog/rtx-pro-5000-72gb-blackwell-gpu/