NVIDIA · Blackwell
Best AI Models for NVIDIA RTX PRO 4500 Blackwell (32 GB)
Runs 27–31B and 35B MoE models well; 70B doesn't fit in 32 GB.
| # | Model | Speed | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 28.6 t/s | 28.6 tok/s | Q4_K_M · 21.7 GB | |
| 4 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 157 t/s | 157 tok/s | Q4_K_M · 16.5 GB | |
| 5 | Q4_K_M · 19.3 GB · 187 t/s | 187 tok/s | Q4_K_M · 19.3 GB |
| # | Model | Speed | Coding Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 28.6 t/s | 28.6 tok/s | Q4_K_M · 21.7 GB | |
| 3 | Qwen3.6 35B A3BAlibaba Q4_K_M · 22.0 GB · 139 t/s | 139 tok/s | Q4_K_M · 22.0 GB | |
| 4 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 157 t/s | 157 tok/s | Q4_K_M · 16.5 GB | |
| 5 | Q4_K_M · 19.3 GB · 187 t/s | 187 tok/s | Q4_K_M · 19.3 GB |
| # | Model | Speed | Reasoning Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 28.6 t/s | 28.6 tok/s | Q4_K_M · 21.7 GB | |
| 3 | Qwen3.5 27BAlibaba Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 4 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 157 t/s | 157 tok/s | Q4_K_M · 16.5 GB | |
| 5 | Q4_K_M · 19.3 GB · 187 t/s | 187 tok/s | Q4_K_M · 19.3 GB |
| # | Model | Speed | Math Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Qwen3.6 35B A3BAlibaba Q4_K_M · 22.0 GB · 139 t/s | 139 tok/s | Q4_K_M · 22.0 GB | |
| 4 | Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 28.6 t/s | 28.6 tok/s | overall | Q4_K_M · 21.7 GB |
| 5 | Q4_K_M · 19.3 GB · 187 t/s | 187 tok/s | Q4_K_M · 19.3 GB |
| # | Model | Speed ↓ | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 4BAlibabaHow to run → Q4_K_M · 3.4 GB · 179 t/s | 179 tok/s | Q4_K_M · 3.4 GB | |
| 2 | Q4_K_M · 19.3 GB · 187 t/s | 187 tok/s | Q4_K_M · 19.3 GB | |
| 3 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 173 t/s | 173 tok/s | Q4_K_M · 13.1 GB | |
| 4 | Qwen3 30B A3B Thinking 2507Alibaba Q4_K_M · 19.4 GB · 162 t/s | 162 tok/s | Q4_K_M · 19.4 GB | |
| 5 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 157 t/s | 157 tok/s | Q4_K_M · 16.5 GB |
| # | Model | Speed | Vision Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 33.4 t/s | 33.4 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Gemma 4 31B ITGoogle Q4_K_M · 21.7 GB · 28.6 t/s | 28.6 tok/s | Q4_K_M · 21.7 GB | |
| 4 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 157 t/s | 157 tok/s | Q4_K_M · 16.5 GB | |
| 5 | MiniCPM V 4.7 35B A3BOpenBMB Q4_K_M · 22.1 GB · 156 t/s | 156 tok/s | Q4_K_M · 22.1 GB |
The RTX PRO 4500 Blackwell runs 27–31B and 35B MoE models well. Blackwell generation, released 2025.
RTX PRO 4500 Blackwell strengths and weaknesses
Strengths
- Runs 35B MoE models: Qwen3.6 35B A3B at 139 tok/s in 22 GB
- Room for long context: Qwen3.8 27B at Q4 uses 17.9 GB, leaving 14.1 GB free
- Reads long prompts fast: A 4K-token prompt on Qwen3.8 27B starts answering in 2.5 s
- FP8 and NVFP4 run natively: Blackwell takes FP8 and NVFP4 checkpoints in vLLM as published
Weaknesses
- 70B models don't fit: Llama 3.3 70B Instruct needs 51.3 GB at Q4 vs 32 GB here
Output speed depends on memory bandwidth (896 GB/s). Grey rows need more than 32 GB.
- 9BQwen3.5 9B · Q4_K_M · 6.6 GB91.6tok/s
- 14BQwen3 14B · Q4_K_M · 11.9 GB61.3tok/s
- 27–31BQwen3.8 27B · Q4_K_M · 17.9 GB33.4tok/s
- 35B MoEQwen3.6 35B A3B · Q4_K_M · 22.0 GB139tok/s
- 70BLlama 3.3 70B Instruct · not at 4-bitneeds 51.3 GB
- 120B MoEGPT OSS 120B · not at 4-bitneeds 70.8 GB
Estimated speeds: Q4_K_M, 4K context, one user. How we estimate
All 1176 models on the RTX PRO 4500 Blackwell
Featured onlyBest quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works
| Model | Quant | VRAM | Speed | Context | Score | Fit |
|---|---|---|---|---|---|---|
| Q4_K_M | 2.3 GB | 348.7 t/s | 4K | S100, Runs great | ||
Gpt2 MiniChat | Q4_K_M | 0.0 GB | 19413.3 t/s | 1K | S100, Runs great | |
Saiga Llama3 8BChat | Q4_K_M | 6.2 GB | 108.1 t/s | 8K | S93, Runs great | |
| BF16 | 12.7 GB | 55.6 t/s | — | A82, Runs well | ||
| Q4_K_M | 7.4 GB | 106.1 t/s | 131K | S92, Runs great | ||
KONI Llama3.1 8B Instruct 20241024ChatFunctions | Q4_K_M | 7.3 GB | 108.1 t/s | 131K | S93, Runs great | |
| Q4_K_M | 4.1 GB | 236.7 t/s | 131K | S100, Runs great | ||
| Q4_K_M | 17.8 GB | 39.2 t/s | 33K | A75, Runs well | ||
Gemma 3 4B ITVision | Q4_K_M | 5 GB | 205.1 t/s | — | S100, Runs great | |
Granite 4.0 H TinyChatFunctions | Q4_K_M | 5.8 GB | 257.8 t/s | 131K | S100, Runs great | |
Tiny Dummy Qwen2Chat | Q4_K_M | 0.5 GB | 1941.3 t/s | 33K | S100, Runs great | |
C4ai Command R 08 2024ChatFunctions | Q4_K_M | 25.1 GB | 27.3 t/s | — | A68, Runs well | |
DeepSeek R1 Distill Qwen 14B Abliterated v2ChatReasoning | Q4_K_M | 12.4 GB | 60.9 t/s | 131K | A84, Runs well | |
Sarvam 1Chat | Q4_K_M | 2.8 GB | 284.1 t/s | 8K | S100, Runs great | |
DeepSWE PreviewChat | Q4_K_M | 22.6 GB | 28.7 t/s | 41K | A69, Runs well |
Rent NVIDIA RTX PRO 4500 Blackwell in the cloud
Hourly rate on demand for one RTX PRO 4500 Blackwell.
Vast.aiMarketplace
from$0.36/hr
Buying one ($2,600) costs as much as ~7,200 hours of rental, about 30 months at 8 hours a day.
Cost per million tokens
Output-token cost on an RTX PRO 4500 Blackwell rented at $0.36/hr, one request at a time. Batch several users on the same card and the cost per token drops several-fold.
| Model | Quant | Speed | Per 1M output |
|---|---|---|---|
| NVIDIA Nemotron 3.5 Lightning 30B A3B BF16 | 187 tok/s | $0.54$0.0052 input | |
| Gemma 4 26B A4B IT | 157 tok/s | $0.64$0.0097 input | |
| Qwen3.6 35B A3B | 139 tok/s | $0.72$0.010 input | |
| Qwen3.8 27B | 33 tok/s | $3.01$0.061 input | |
| Muse Glimmer 30B | 32 tok/s | $3.16$0.066 input |
Specs
- Architecture
- Blackwell
- VRAM
- 32 GB GDDR7
- Memory bandwidth
- 896.0 GB/s
- FP16 (dense)
- 203.00 TFLOPS
- CUDA cores
- 10,496
- Tensor cores
- 328
- Compute capability
- 12.0
- TDP
- 200 W
- Released
- 2025-03-18
- MSRP
- $2,600
RTX PRO 4500 Blackwell FAQ
What is the largest model the RTX PRO 4500 Blackwell can run?
Qwen3.6 35B A3B (36.0B) is the largest at 4-bit or better: at Q4_K_M it uses 22.0 GB of 32 GB, and runs at about 139 tok/s.
Can the RTX PRO 4500 Blackwell run Qwen3.8 27B?
Yes. Qwen3.8 27B fits at Q4_K_M, using 17.9 GB of 32 GB, and it is the best model that runs well on this card.
Can the RTX PRO 4500 Blackwell run 70B models?
No. Llama 3.3 70B Instruct needs 51.3 GB at Q4, more than the 32 GB on this card.
How many tokens per second does the RTX PRO 4500 Blackwell give?
About 33.4 tok/s on Qwen3.8 27B at Q4_K_M, with the first token after 2.5 s on a 4K-token prompt.
What power supply does the RTX PRO 4500 Blackwell need?
Plan a 400 W PSU (2× the 200 W TDP).
Is 32 GB enough for coding and agents?
Yes for coding: Qwen3.8 27B, the best of the 10 coding models that fit, runs at about 33 tok/s. For agents, not yet: no agent-scored model fits in less than 96 GB.