NVIDIA · Blackwell · 24 GB VRAM tier
Best AI Models for NVIDIA RTX PRO 4000 Blackwell SFF Edition (24 GB)
Runs 27–31B and 35B MoE models well; 70B doesn't fit in 24 GB.
24 GB GDDR7432 GB/s70 WBlackwell 2025
70 W low-profile dual-slot card for small-form-factor builds — 432 GB/s, well below the 145 W RTX PRO 4000's bandwidth. NVIDIA published only "Q4 2025" availability, so the release date here is the quarter, not a confirmed retail day.
| # | Model | Speed | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 tok/s | Q4_K_M · 16.5 GB | |
| 4 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 94.8 t/s | 94.8 tok/s | Q4_K_M · 13.1 GB | |
| 5 | Q4_K_M · 19.3 GB · 125 t/s | 125 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 · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 35B A3BAlibaba Q4_K_M · 22.0 GB · 79.9 t/s | 79.9 tok/s | Q4_K_M · 22.0 GB | |
| 3 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 tok/s | Q4_K_M · 16.5 GB | |
| 4 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 94.8 t/s | 94.8 tok/s | Q4_K_M · 13.1 GB | |
| 5 | Q4_K_M · 19.3 GB · 125 t/s | 125 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 · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.5 27BAlibaba Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 tok/s | Q4_K_M · 16.5 GB | |
| 4 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 94.8 t/s | 94.8 tok/s | Q4_K_M · 13.1 GB | |
| 5 | Q4_K_M · 19.3 GB · 125 t/s | 125 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 · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Qwen3.6 35B A3BAlibaba Q4_K_M · 22.0 GB · 79.9 t/s | 79.9 tok/s | Q4_K_M · 22.0 GB | |
| 4 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 tok/s | overall | Q4_K_M · 16.5 GB |
| 5 | Q4_K_M · 19.3 GB · 125 t/s | 125 tok/s | Q4_K_M · 19.3 GB |
| # | Model | Speed | Long context Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.8 27BAlibabaHow to run → Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 tok/s | Q4_K_M · 16.5 GB | |
| 4 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 94.8 t/s | 94.8 tok/s | Q4_K_M · 13.1 GB | |
| 5 | LFM2.5 24B A2BLiquid AI Q4_K_M · 15.0 GB · 131 t/s | 131 tok/s | Q4_K_M · 15.0 GB |
| # | Model | Speed ↓ | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Q4_K_M · 19.4 GB · 100 t/s | 100 tok/s | Q4_K_M · 19.4 GB | |
| 2 | Q4_K_M · 19.3 GB · 125 t/s | 125 tok/s | Q4_K_M · 19.3 GB | |
| 3 | Qwen3 30B A3B Instruct 2507Alibaba Q4_K_M · 19.4 GB · 100 t/s | 100 tok/s | Q4_K_M · 19.4 GB | |
| 4 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 94.8 t/s | 94.8 tok/s | Q4_K_M · 13.1 GB | |
| 5 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 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 · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 2 | Qwen3.6 27BAlibaba Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB | |
| 3 | Gemma 4 26B A4B ITGoogle Q4_K_M · 16.5 GB · 87.5 t/s | 87.5 tok/s | Q4_K_M · 16.5 GB | |
| 4 | Gemma 3 27B ITGoogle Q4_K_M · 21.7 GB · 15.5 t/s | 15.5 tok/s | Q4_K_M · 21.7 GB | |
| 5 | JEV 27B VLautotrust Q4_K_M · 17.9 GB · 16.1 t/s | 16.1 tok/s | Q4_K_M · 17.9 GB |
See all 327 models on the RTX PRO 4000 Blackwell SFF Edition →Best quant that fits · speeds estimated on this card
The RTX PRO 4000 Blackwell SFF Edition runs 27–31B and 35B MoE models well. Blackwell generation, released 2025.
GenerationBlackwellreleased 2025
RTX PRO 4000 Blackwell SFF Edition strengths and weaknesses
Strengths
- Runs 35B MoE models: Qwen3.6 35B A3B at 79.9 tok/s in 22 GB
- Room for long context: Qwen3.8 27B at Q4 uses 17.9 GB, leaving 6.1 GB free
- Low power draw: 70 W board power; a 150 W PSU is enough
Weaknesses
- 70B models don't fit: Llama 3.3 70B Instruct needs 51.3 GB at Q4 vs 24 GB here
Output speed depends on memory bandwidth (432 GB/s). Grey rows need more than 24 GB.
- 9BQwen3.5 9B · Q4_K_M · 6.6 GB44.2tok/s
- 14BQwen3 14B · Q4_K_M · 11.9 GB29.6tok/s
- 27–31BQwen3.8 27B · Q4_K_M · 17.9 GB16.1tok/s
- 35B MoEQwen3.6 35B A3B · Q4_K_M · 22.0 GB79.9tok/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 1141 models on the RTX PRO 4000 Blackwell SFF Edition
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 |
|---|---|---|---|---|---|---|
Qwen AgentWorld 35B A3BChatFunctions | Q4_K_M | 21.3 GB | 102.9 t/s | 262K | A79, Runs well | |
| BF16 | 2.7 GB | 115.1 t/s | 41K | S94, Runs great | ||
Qwen3 1.7BChatFunctions | Q4_K_M | 3.4 GB | 160.5 t/s | 41K | S97, Runs great | |
| Q4_K_M | 21.5 GB | 146.2 t/s | 262K | A83, Runs well | ||
Opt 125MChat | BF16 | 0.3 GB | 1002.9 t/s | 2K | S100, Runs great | |
SmolVLM 500M InstructVision | Q4_K_M | 0.9 GB | 407.0 t/s | 8K | S100, Runs great | |
Moore LmChat | BF16 | 0.7 GB | 3120.0 t/s | — | S100, Runs great | |
LFM2.5 VL 450MVisionFunctions | Q4_K_M | 0.8 GB | 438.8 t/s | 128K | S100, Runs great | |
Seed OSS 36B InstructChatFunctions | Q4_K_S | 22.8 GB | 13.4 t/s | 524K | C30, Usable | |
| Q4_K_M | 3.5 GB | 81.9 t/s | 131K | S88, Runs great | ||
| Q4_K_M | 0.0 GB | 28080.0 t/s | 1K | S100, Runs great | ||
| Q4_K_M | 0.0 GB | 28080.0 t/s | 1K | S100, Runs great | ||
GPT S 5MChat | BF16 | 0.3 GB | 877.5 t/s | 1K | S100, Runs great | |
Olmo 3.1 32B InstructChatFunctions | Q4_K_M | 21.5 GB | 13.9 t/s | 66K | C39, Usable | |
SKT ST X 0 3BChat | Q4_K_M | 2.4 GB | 170.3 t/s | 2K | S98, Runs great |
Specs
- Architecture
- Blackwell
- VRAM
- 24 GB GDDR7
- Memory bandwidth
- 432.0 GB/s
- CUDA cores
- 8,960
- Tensor cores
- 280
- Compute capability
- 12.0
- TDP
- 70 W
- Released
- 2025-10-01
RTX PRO 4000 Blackwell SFF Edition FAQ
What is the largest model the RTX PRO 4000 Blackwell SFF Edition 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 24 GB, and runs at about 79.9 tok/s.
Can the RTX PRO 4000 Blackwell SFF Edition run Qwen3.8 27B?
Yes. Qwen3.8 27B fits at Q4_K_M, using 17.9 GB of 24 GB, and it is the best model that runs well on this card.
Can the RTX PRO 4000 Blackwell SFF Edition run 70B models?
No. Llama 3.3 70B Instruct needs 51.3 GB at Q4, more than the 24 GB on this card.
How many tokens per second does the RTX PRO 4000 Blackwell SFF Edition give?
About 16.1 tok/s on Qwen3.8 27B at Q4_K_M.
What power supply does the RTX PRO 4000 Blackwell SFF Edition need?
Plan a 150 W PSU (2× the 70 W TDP).
Is 24 GB enough for coding and agents?
Yes for coding: Qwen3.8 27B, the best of the 8 coding models that fit, runs at about 16 tok/s. For agents, not yet: no agent-scored model fits in less than 96 GB.