RTX PRO 6000 Blackwell Workstation Edition: 96 GB, what it runs and what it costs
Specs checked · Rental prices from 18m ago · Aliteq Team · How we compute this
The short answer
The NVIDIA RTX PRO 6000 Blackwell Workstation Edition has 96 GB of VRAM and 1792 GB/s of memory bandwidth (2025). At Q4_K_M quality or better with an 8K context it runs 40 of the open models we track; the largest that fits with room to spare is Qwen3.5 122B-A10B (MoE). The cheapest rental right now is $0.520/hr (Vast.ai, interruptible).
- VRAM
- 96 GB
- Bandwidth
- 1792 GB/s
- Board power
- 600 W
- Launched
- 2025
Source: NVIDIA specification · bandwidth not yet confirmed by a second source, so we don't estimate speed from it
Buy it or rent it
| Provider | Type | From, per hour | Rent |
|---|---|---|---|
| Vast.ai | Spot (can be interrupted) | $0.520/hr | Vast.ai |
| Vast.ai | On demand | $1.134/hr | Vast.ai |
| Lium | Spot (can be interrupted) | $1.590/hr | Lium |
| Lium | On demand | $1.850/hr | Lium |
| Runpod | On demand | $2.590/hr | Runpod |
Lowest price per provider in our hourly snapshot (18m ago). Rent buttons are referral links. Every card side by side: cloud GPU prices.
Which AI models run on the RTX PRO 6000 Blackwell Workstation Edition
Computed by our VRAM engine for an 8K context: the best quality each model fits at in 96 GB. Speeds are upper-bound estimates from memory bandwidth, not benchmarks.
| Model | Best fit | VRAM used | ~tok/s |
|---|---|---|---|
| Qwen3.5 122B-A10B (MoE) | Q5_K_M | 84 GB87% | — |
| NVIDIA Nemotron 3 Super 120B-A12B (MoE) | Q6_K | 95 GB99% · tight | — |
| gpt-oss 120B (MoE) | Q6_K | 91 GB94% · tight | — |
| Llama 4 Scout 17B-16E (MoE) | Q6_K | 84 GB88% | — |
| GLM-4.5-Air (MoE) | Q6_K | 83 GB87% | — |
| Qwen3-Next 80B-A3B (MoE) | Q8_0 | 80 GB83% | — |
| Qwen3-Coder-Next 80B-A3B (MoE) | Q8_0 | 80 GB83% | — |
| DeepSeek R1 Distill 70B | Q8_0 | 73 GB76% | — |
| Mixtral 8x7B (MoE) | F16 | 89 GB92% · tight | — |
| Qwen3.6 35B-A3B (MoE) | F16 | 68 GB71% | — |
| Ornith-1.5 35B-A3B (MoE) | F16 | 68 GB71% | — |
| Qwen3.5 35B-A3B (MoE) | F16 | 68 GB71% | — |
Fits only below Q4 quality: Qwen3 235B-A22B (MoE) (Q2_K), MiniMax-M2.5 (MoE) (Q2_K), Qwen3.8-Flash-Next (MoE) (IQ4_XS).
Too big for 96 GB (rent a bigger card instead): what it costs to run MiniMax-M3 (MoE), what it costs to run GLM-5.3 (MoE), what it costs to run DeepSeek V4.1 Flash (MoE), what it costs to run DeepSeek V4 Pro (MoE).
Check a model and context of your own in the VRAM calculator.
AI video on the RTX PRO 6000 Blackwell Workstation Edition
Alibaba's Wan 2.2 is an open video model you can run yourself in ComfyUI. With 96 GB the RTX PRO 6000 Blackwell Workstation Edition fits:
- Wan 2.2 14B at 720p without offloading (80 GB and up, per Runpod's Wan 2.2 guide)
- Wan 2.2 14B in fp16, or in fp8 at 720p (48 GB and up, per Runpod's Wan 2.2 guide)
- Wan 2.2 14B at 640x640 with fp8 weights (24 GB and up, per Runpod's Wan 2.2 guide)
- Wan 2.2 5B, with ComfyUI's native offloading (8 GB and up, per ComfyUI's Wan 2.2 docs)
What each setup needs, the cheapest card to rent and what one clip costs: cloud GPUs for ComfyUI with Flux and Wan 2.2. Against an app or an API: AI video cost per clip, app vs API vs rented GPU. Rather not run a model yourself? Hosted video models such as Seedance, Veo and Hailuo are compared in Kling AI alternatives, priced per clip.
Cards to compare it with
- RTX PRO 6000 Blackwell Max-Q Workstation Edition · 96 GB
- Apple M2 Max · 96 GB
- Apple M4 Max (16-core CPU / 40-core GPU) · 128 GB
- DGX Spark (GB10 Grace Blackwell) · 128 GB
- Apple M4 Pro · 64 GB
- AMD Ryzen AI Max+ 395 (Strix Halo,) · 128 GB
Every card by VRAM tier: all GPUs for gaming and local AI.