RUNLOCALAIv38
->Will it run?Best GPUCompareTroubleshootStartLearnPulseModelsHardwareToolsBench
Run check
RUNLOCALAI

Independently operated catalog for local-AI hardware and software. Hand-written verdicts. Source-cited claims. Reproducible commands when we have them.

OP·Fredoline Eruo
DIR
  • Models
  • Hardware
  • Tools
  • Benchmarks
TOOLS
  • Will it run?
  • Compare hardware
  • Cost vs cloud
  • Choose my GPU
  • Prompting kits
  • Quick answers
REF
  • All buyer guides
  • Learn local AI
  • Methodology
  • Glossary
  • Errors KB
  • Trust
EDITOR
  • About
  • Author
  • How we make money
  • Editorial policy
  • Contact
LEGAL
  • Privacy
  • Terms
  • Sitemap
MAIL · MONTHLY DIGEST
Get monthly local AI changes
Monthly recap. No spam.
DISCLOSURE

Some links on this site are affiliate links (Amazon Associates and other first-class retailers). When you buy through them, we earn a small commission at no extra cost to you. Affiliate links do not influence our verdicts — there are cards we rate highly that we don't have affiliate relationships with, and cards that sell well that we refuse to recommend. Read more →

© 2026 runlocalai.coIndependently operated
RUNLOCALAI · v38
Will it run? / Apple Mac Studio (M3 Ultra) / long context

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

Build: Apple Mac Studio (M3 Ultra) + — + 32 GB RAM (windows)

Memory: 32 GB unified memory
Runner: llama.cpp (Metal)
AnyChatCodingAgentsReasoningVisionLong contextCreative

Runs comfortably
138 models

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

#1Phi-3.5 Mini Instruct
3.8B
phi
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 6.8 GBHeadroom: 17.2 GB
ollama run phi3.5:3.8b
123
tok/s
Estimated
Weights
4.04 GB
KV cache
1.90 GB
Activations
0.21 GB
Runtime
0.70 GB
Model details →
#2Ministral 3B Instruct
3B
mistral
Quant: Q4_K_MContext: 8,192VRAM: 4.1 GBHeadroom: 19.9 GB
274
tok/s
Estimated
Weights
1.81 GB
KV cache
1.50 GB
Activations
0.10 GB
Runtime
0.70 GB
Model details →
#3Codestral Mamba 7B
7B
mistral
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.6 GBHeadroom: 15.4 GB
117
tok/s
Estimated
Weights
4.23 GB
KV cache
3.50 GB
Activations
0.22 GB
Runtime
0.70 GB
Model details →
#4Falcon Mamba 7B
7B
falcon
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.6 GBHeadroom: 15.4 GB
117
tok/s
Estimated
Weights
4.23 GB
KV cache
3.50 GB
Activations
0.22 GB
Runtime
0.70 GB
Model details →
#5Gemma 4 E4B (Effective 4B)
4B
gemma
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 7.2 GBHeadroom: 16.8 GB
ollama run gemma4:e4b
117
tok/s
Estimated
Weights
4.25 GB
KV cache
2.00 GB
Activations
0.22 GB
Runtime
0.70 GB
Model details →
#6Command R7B (12-2024)
8B
command-r
Quant: Q4_K_MContext: 8,192VRAM: 9.8 GBHeadroom: 14.2 GB
103
tok/s
Estimated
Weights
4.83 GB
KV cache
4.00 GB
Activations
0.25 GB
Runtime
0.70 GB
Model details →
#7Ministral 8B Instruct
8B
mistral
Quant: Q4_K_MContext: 8,192VRAM: 9.8 GBHeadroom: 14.2 GB
103
tok/s
Estimated
Weights
4.83 GB
KV cache
4.00 GB
Activations
0.25 GB
Runtime
0.70 GB
Model details →
#8Qwen 2.5 7B Instruct
7B
qwen
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 9.0 GBHeadroom: 15.0 GB
ollama run qwen2.5:7b
67
tok/s
Estimated
Weights
7.44 GB
KV cache
0.47 GB
Activations
0.38 GB
Runtime
0.70 GB
Model details →
#9Qwen 3 8B
8B
qwen
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 13.6 GBHeadroom: 10.4 GB
ollama run qwen3:8b
58
tok/s
Estimated
Weights
8.50 GB
KV cache
4.00 GB
Activations
0.43 GB
Runtime
0.70 GB
Model details →
#10InternLM 2.5 7B Chat
7B
internlm
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 8.6 GBHeadroom: 15.4 GB
117
tok/s
Estimated
Weights
4.23 GB
KV cache
3.50 GB
Activations
0.22 GB
Runtime
0.70 GB
Model details →
#11DeepSeek V2 Lite Chat
15.7B
deepseek
Commercial OK
Quant: Q4_K_MContext: 8,192VRAM: 18.5 GBHeadroom: 5.5 GB
342
tok/s
Estimated
Weights
9.48 GB
KV cache
7.85 GB
Activations
0.48 GB
Runtime
0.70 GB
Model details →
#12ColPali v1.3
3B
gemma
Commercial OK
Quant: Q4_K_MContext: 0VRAM: 2.6 GBHeadroom: 21.4 GB
274
tok/s
Estimated
Weights
1.81 GB
KV cache
0.00 GB
Activations
0.09 GB
Runtime
0.70 GB
Model details →

Runs with tradeoffs
20 models

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

Nemotron 3 Nano (30B-A3B)
30B
other
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 23.5 GBHeadroom: 0.5 GB
  • • Tight VRAM fit — only 0.5 GB headroom left for context growth
ollama run nemotron3:nano
27
tok/s
Estimated
Weights
18.11 GB
KV cache
3.75 GB
Activations
0.91 GB
Runtime
0.70 GB
Model details →
Mistral Nemo 12B Instruct
12B
mistral
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 20.1 GBHeadroom: 3.9 GB
  • • Tight VRAM fit — only 3.9 GB headroom left for context growth
ollama run mistral-nemo:12b
39
tok/s
Estimated
Weights
12.75 GB
KV cache
6.00 GB
Activations
0.65 GB
Runtime
0.70 GB
Model details →
Qwen 3 14B
14B
qwen
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 23.3 GBHeadroom: 0.7 GB
  • • Tight VRAM fit — only 0.7 GB headroom left for context growth
ollama run qwen3:14b
33
tok/s
Estimated
Weights
14.88 GB
KV cache
7.00 GB
Activations
0.75 GB
Runtime
0.70 GB
Model details →
Qwen 2.5 14B Instruct
14B
qwen
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 23.3 GBHeadroom: 0.7 GB
  • • Tight VRAM fit — only 0.7 GB headroom left for context growth
ollama run qwen2.5:14b
33
tok/s
Estimated
Weights
14.88 GB
KV cache
7.00 GB
Activations
0.75 GB
Runtime
0.70 GB
Model details →
Gemma 4 Turkish 26B (4B active)
26B
gemma
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 20.4 GBHeadroom: 3.6 GB
  • • Tight VRAM fit — only 3.6 GB headroom left for context growth
32
tok/s
Estimated
Weights
15.70 GB
KV cache
3.25 GB
Activations
0.79 GB
Runtime
0.70 GB
Model details →
Gemma 3 27B
27B
gemma
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 21.2 GBHeadroom: 2.8 GB
  • • Tight VRAM fit — only 2.8 GB headroom left for context growth
ollama run gemma3:27b
30
tok/s
Estimated
Weights
16.30 GB
KV cache
3.38 GB
Activations
0.82 GB
Runtime
0.70 GB
Model details →
Qwen 3 30B-A3B
30B
qwen
Commercial OK
Quant: Q4_K_MContext: 2,048VRAM: 23.5 GBHeadroom: 0.5 GB
  • • Tight VRAM fit — only 0.5 GB headroom left for context growth
ollama run qwen3:30b
27
tok/s
Estimated
Weights
18.11 GB
KV cache
3.75 GB
Activations
0.91 GB
Runtime
0.70 GB
Model details →
Phi-4 14B
14B
phi
Commercial OK
Quant: Q8_0Context: 8,192VRAM: 23.3 GBHeadroom: 0.7 GB
  • • Tight VRAM fit — only 0.7 GB headroom left for context growth
ollama run phi4:14b
33
tok/s
Estimated
Weights
14.88 GB
KV cache
7.00 GB
Activations
0.75 GB
Runtime
0.70 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 24 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 24 GB available after OS overhead.

—
Qwen 3 235B-A22B
235B
qwen
Commercial OK

Needs ~160 GB unified memory minimum at smallest quant; you have 24 GB available after OS overhead.

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

Needs ~420 GB unified memory minimum at smallest quant; you have 24 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 24 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.