What can NVIDIA GeForce RTX 5070 Ti run for vision?

Build: RTX 5070 Ti + Ryzen 7 9700X + 32GB DDR5

Memory: 16 GB VRAM + 32 GB system RAM
Runner: llama.cpp / Ollama (CUDA)

Runs comfortably
14 models

Ranked by fit for vision use case + predicted speed. Click a row for VRAM breakdown.

#1Gemma 4 E4B (Effective 4B)
4B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 8.4 GBHeadroom: 7.6 GB
ollama run gemma4:e4b
137
tok/s
Estimated
Weights
4.40 GB
KV cache
2.00 GB
Activations
0.23 GB
Runtime
1.80 GB
#2Gemma 3 4B
4B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 8.4 GBHeadroom: 7.6 GB
ollama run gemma3:4b
137
tok/s
Estimated
Weights
4.40 GB
KV cache
2.00 GB
Activations
0.23 GB
Runtime
1.80 GB
#3Gemma 4 E2B (Effective 2B)
2B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 5.1 GBHeadroom: 10.9 GB
ollama run gemma4:e2b
274
tok/s
Estimated
Weights
2.20 GB
KV cache
1.00 GB
Activations
0.12 GB
Runtime
1.80 GB
#4Phi-3.5 Vision
4.2B
phi
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 6.5 GBHeadroom: 9.5 GB
230
tok/s
Estimated
Weights
2.50 GB
KV cache
2.10 GB
Activations
0.13 GB
Runtime
1.80 GB
#5LLaVA 1.6 Mistral 7B
7B
other
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 10.0 GBHeadroom: 6.0 GB
138
tok/s
Estimated
Weights
4.50 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
1.80 GB
#6LLaVA-OneVision 7B
7B
other
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 10.0 GBHeadroom: 6.0 GB
138
tok/s
Estimated
Weights
4.50 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
1.80 GB
#7Moondream 2
1.9B
other
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 3.3 GBHeadroom: 12.7 GB
508
tok/s
Estimated
Weights
1.20 GB
KV cache
0.24 GB
Activations
0.06 GB
Runtime
1.80 GB
#8Qwen 2-VL 7B
7B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 10.1 GBHeadroom: 5.9 GB
138
tok/s
Estimated
Weights
4.60 GB
KV cache
3.50 GB
Activations
0.24 GB
Runtime
1.80 GB
#9MiniCPM-V 3 8B
8B
minicpm
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 11.1 GBHeadroom: 4.9 GB
121
tok/s
Estimated
Weights
5.00 GB
KV cache
4.00 GB
Activations
0.26 GB
Runtime
1.80 GB
#10MiniCPM-V 2.6 8B
8B
minicpm
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 11.1 GBHeadroom: 4.9 GB
121
tok/s
Estimated
Weights
5.00 GB
KV cache
4.00 GB
Activations
0.26 GB
Runtime
1.80 GB
#11Molmo 7B-D
8B
other
Commercial OK
Quant: Q4_K_MContext: 4,096VRAM: 9.3 GBHeadroom: 6.7 GB
121
tok/s
Estimated
Weights
5.20 GB
KV cache
2.00 GB
Activations
0.26 GB
Runtime
1.80 GB
#12Qwen 2.5-VL 7B
7B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 9.9 GBHeadroom: 6.1 GB
138
tok/s
Estimated
Weights
4.40 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
1.80 GB

Runs with tradeoffs
17 models

Tight VRAM, partial CPU offload, or context-limited.

Qwen3.5 9B
9B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 13.2 GBHeadroom: 2.8 GB
  • Tight VRAM fit — only 2.8 GB headroom left for context growth
ollama run qwen3.5:9b
107
tok/s
Estimated
Weights
6.60 GB
KV cache
4.50 GB
Activations
0.34 GB
Runtime
1.80 GB
Llama 3.2 11B Vision Instruct
11B
llama
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 15.6 GBHeadroom: 0.4 GB
  • Tight VRAM fit — only 0.4 GB headroom left for context growth
ollama run llama3.2-vision:11b
88
tok/s
Estimated
Weights
7.90 GB
KV cache
5.50 GB
Activations
0.40 GB
Runtime
1.80 GB
Llama 3.2 11B Vision
11B
llama
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 14.1 GBHeadroom: 1.9 GB
  • Tight VRAM fit — only 1.9 GB headroom left for context growth
88
tok/s
Estimated
Weights
6.50 GB
KV cache
5.50 GB
Activations
0.33 GB
Runtime
1.80 GB
Gemma 4 12B
12B
gemma
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 15.8 GBHeadroom: 0.2 GB
  • Tight VRAM fit — only 0.2 GB headroom left for context growth
ollama run gemma4:12b
80
tok/s
Estimated
Weights
7.60 GB
KV cache
6.00 GB
Activations
0.39 GB
Runtime
1.80 GB
Gemma 3 12B
12B
gemma
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 15.5 GBHeadroom: 0.5 GB
  • Tight VRAM fit — only 0.5 GB headroom left for context growth
ollama run gemma3:12b
80
tok/s
Estimated
Weights
7.30 GB
KV cache
6.00 GB
Activations
0.37 GB
Runtime
1.80 GB
Pixtral 12B
12B
mistral
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 15.2 GBHeadroom: 0.8 GB
  • Tight VRAM fit — only 0.8 GB headroom left for context growth
ollama run pixtral:12b
80
tok/s
Estimated
Weights
7.00 GB
KV cache
6.00 GB
Activations
0.36 GB
Runtime
1.80 GB
GLM-4V 9B
13.9B
glm
Quant: Q4_K_MContext: 2,048VRAM: 12.5 GBHeadroom: 3.5 GB
  • Tight VRAM fit — only 3.5 GB headroom left for context growth
69
tok/s
Estimated
Weights
8.50 GB
KV cache
1.74 GB
Activations
0.43 GB
Runtime
1.80 GB
Ministral 3 14B
14B
mistral
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 13.1 GBHeadroom: 2.9 GB
  • Tight VRAM fit — only 2.9 GB headroom left for context growth
ollama run ministral-3:14b
69
tok/s
Estimated
Weights
9.10 GB
KV cache
1.75 GB
Activations
0.46 GB
Runtime
1.80 GB

What if you upgraded?

Hypothetical scenarios. We re-ran the compatibility engine for each.

+32 GB system RAM

~$80–150

Doubles your CPU-offload working set. Helps when models don't quite fit in VRAM.

Unlocks: 128 new comfortable, 112 new tradeoff

  • Qwen 3 0.6B
  • Llama 3.1 8B Instruct
  • Qwen 3 8B
  • Llama 3.2 3B Instruct

Upgrade to NVIDIA RTX 2080 Ti 22GB (China-mod)

~$350

22 GB VRAM (vs your 16 GB) plus a bandwidth jump from ~896 GB/s to ~616 GB/s.

Unlocks: 155 new comfortable

  • Qwen 3 0.6B
  • Llama 3.1 8B Instruct
  • Qwen 3 8B
  • Gemma 4 12B

Add a second NVIDIA GeForce RTX 5070 Ti

~$849

Tensor parallelism splits the model across both cards, effectively doubling VRAM. Bandwidth doesn't double — runs ~1.5× the single-card speed in practice.

Unlocks: 202 new comfortable

  • Qwen 3 0.6B
  • Llama 3.1 8B Instruct
  • Qwen 3 30B-A3B
  • Qwen 2.5 Coder 32B Instruct

Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.

Won't run
top 5 popular models

Need more memory than you have. Shown for orientation.

DeepSeek V4 Pro (1.6T MoE)
1600B
deepseek
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

Qwen 3.5 235B-A17B (MoE)
397B
qwen
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

Qwen 3 235B-A22B
235B
qwen
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

DeepSeek R1 (671B reasoning)
671B
deepseek
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

Llama 4 Scout
109B
llama
Commercial OK

Even with CPU offload, needs more memory than your VRAM (16 GB) + 60% of system RAM (19 GB) combined.

How to read these numbers

Measured here
Measured here - RunLocalAI ran this exact combo on owner hardware with public evidence.

Source-backed
Source-backed / community - a reproduced public source supports the speed, but it is not labeled as owner-measured.

Extrapolated
Extrapolated - predicted from a measured benchmark on similar-bandwidth hardware.

Estimated
Estimated - formula based on VRAM bandwidth and model architecture; not a benchmark row.

RunLocalAI Will-It-Run Framework →

Want a specific benchmark we don't have? Email Contact support and we'll prioritize it.