Skip to content

Olmo 3.1 32B on M3 Ultra 96 GB: VRAM, speed and fit

Olmo 3.1 32B at 8-bit runs on the M3 Ultra 96 GB with 64K of context at about 15 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on M3 Ultra 96 GB · 8-bit (Q8_0) · Ollama / LM Studio · Chat

GPU memory
61.2 GB
of 96.0 GB
Context
64K
auto
Speed
15 tok/s
Reading speed
First reply
6.2 s
2K prompt

Where the memory goes

Model
31.9 GB
Context
4.8 GB
Overhead
612 MB
OS
24.0 GB
Free
34.8 GB
Memory
61.2 GB of 96.0 GB
Context
64K tokens · auto
Speed
15 tok/s · reading speed
First reply
6.2 s for a 2K prompt
Running cost
$3.55/mo at home · $61/mo on a cloud 2× RTX A4500

Validation: Not validated

What Olmo 3.1 32B on M3 Ultra 96 GB was checked against
CheckTargetResultStatus
Fits in GPU memory72.0 GB usable37.2 GB neededPass
Headroom5% free34.8 GB free (48%)Pass
Context for chat16K or more64KPass
Writing speed15 tok/s or more15 tok/sCheck
First replyunder 3 s, 2K prompt6.2 sFails

Targets: a reply you can read along with, a first reply the use case tolerates, and 5% of memory left free. Calculated from published specs, not benchmarked.

The setup

Model
Olmo 3.1 32B
Parameters
32.2B
Precision
8-bit (Q8_0)
Context
64K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
M3 Ultra 96 GB
GPU can use
72.0 GB of 96.0 GB

Running cost

At home
$3.55/mo
162 W at the wall
Renting instead
$61/mo
2× RTX A4500 at $0.50/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 21, 2026.

How to run it

Ollama
ollama run olmo-3.1:32b-instruct
llama.cpp
llama-server -hf unsloth/Olmo-3.1-32B-Instruct-GGUF
vLLM
vllm serve allenai/Olmo-3.1-32B-Instruct
MLX (Mac)
mlx_lm.chat --model mlx-community/Olmo-3.1-32B-Instruct-4bit

Olmo 3.1 32B on similar machines

All machines: Olmo 3.1 32B VRAM requirements. Everything that runs here: what can the M3 Ultra 96 GB run?

Other models that run on M3 Ultra 96 GB

Questions

Can Olmo 3.1 32B run on M3 Ultra 96 GB?

Olmo 3.1 32B at 8-bit runs on the M3 Ultra 96 GB with 64K of context at about 15 tokens a second.

Report VR-133GKFY, calculated Nov 22, 2026 from published specs, not benchmarked.