What can Apple Mac Studio (M3 Ultra) run for vision?

Build: Apple Mac Studio (M3 Ultra) + — + 192 GB RAM (macos)

Memory: 192 GB unified memory
Runner: MLX-LM (Apple Metal)
AnyChatCodingAgentsReasoningVisionLong contextCreative

Runs comfortably
37 models

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

#1Gemma 4 E2B (Effective 2B)
2B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 3.8 GBHeadroom: 180.2 GB
ollama run gemma4:e2b
207
tok/s
Estimated
Weights
2.20 GB
KV cache
1.00 GB
Activations
0.12 GB
Runtime
0.50 GB
Model details →
#2Phi-3.5 Vision
4.2B
phi
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 5.2 GBHeadroom: 178.8 GB
174
tok/s
Estimated
Weights
2.50 GB
KV cache
2.10 GB
Activations
0.13 GB
Runtime
0.50 GB
Model details →
#3Moondream 2
1.9B
other
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 2.0 GBHeadroom: 182.0 GB
384
tok/s
Estimated
Weights
1.20 GB
KV cache
0.24 GB
Activations
0.06 GB
Runtime
0.50 GB
Model details →
#4Qwen 2.5-VL 3B
3B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 4.1 GBHeadroom: 179.9 GB
243
tok/s
Estimated
Weights
2.00 GB
KV cache
1.50 GB
Activations
0.11 GB
Runtime
0.50 GB
Model details →
#5LLaVA 1.6 Mistral 7B
7B
other
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.7 GBHeadroom: 175.3 GB
104
tok/s
Estimated
Weights
4.50 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
0.50 GB
Model details →
#6LLaVA-OneVision 7B
7B
other
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.7 GBHeadroom: 175.3 GB
104
tok/s
Estimated
Weights
4.50 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
0.50 GB
Model details →
#7Qwen 2-VL 7B
7B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.8 GBHeadroom: 175.2 GB
104
tok/s
Estimated
Weights
4.60 GB
KV cache
3.50 GB
Activations
0.24 GB
Runtime
0.50 GB
Model details →
#8Qwen 2.5-VL 7B
7B
qwen
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.6 GBHeadroom: 175.4 GB
104
tok/s
Estimated
Weights
4.40 GB
KV cache
3.50 GB
Activations
0.23 GB
Runtime
0.50 GB
Model details →
#9Gemma 4 E4B (Effective 4B)
4B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 7.1 GBHeadroom: 176.9 GB
ollama run gemma4:e4b
104
tok/s
Estimated
Weights
4.40 GB
KV cache
2.00 GB
Activations
0.23 GB
Runtime
0.50 GB
Model details →
#10Gemma 3 4B
4B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 7.1 GBHeadroom: 176.9 GB
ollama run gemma3:4b
104
tok/s
Estimated
Weights
4.40 GB
KV cache
2.00 GB
Activations
0.23 GB
Runtime
0.50 GB
Model details →
#11MiniCPM-V 3 8B
8B
minicpm
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 9.8 GBHeadroom: 174.2 GB
91
tok/s
Estimated
Weights
5.00 GB
KV cache
4.00 GB
Activations
0.26 GB
Runtime
0.50 GB
Model details →
#12MiniCPM-V 2.6 8B
8B
minicpm
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 9.8 GBHeadroom: 174.2 GB
91
tok/s
Estimated
Weights
5.00 GB
KV cache
4.00 GB
Activations
0.26 GB
Runtime
0.50 GB
Model details →

What if you upgraded?

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

Move up an Apple memory tier

~$200–400 over base

On Apple Silicon, more unified memory is the only path forward — VRAM and system RAM are the same pool.

Shop this upgrade↗

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

Needs ~1024 GB unified memory minimum at smallest quant; you have 184 GB available after OS overhead.

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

Needs ~256 GB unified memory minimum at smallest quant; you have 184 GB available after OS overhead.

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

Needs ~420 GB unified memory minimum at smallest quant; you have 184 GB available after OS overhead.

—
DeepSeek V4 Flash (284B MoE)
284B
deepseek
Commercial OK

Needs ~192 GB unified memory minimum at smallest quant; you have 184 GB available after OS overhead.

—
GLM-5.2
753B
glm
Commercial OK

Needs ~528 GB unified memory minimum at smallest quant; you have 184 GB available after OS overhead.

—

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.