Skip to content
Local AI Wizard
Your machine
Theme

Olmo 3 family: which size to run

2 sizes, 7B–32B parameters · Ai2

Checked

Sizes

Olmo 3 sizes
ModelArchitecture
Olmo 3 7B InstructAi2 · Apache 2.0 · Nov 20, 20257BDense65,536No LiveBench result for this model in release 2026-06-251484.54.1
Olmo 3 32B ThinkAi2 · Apache 2.0 · Nov 20, 202532BDense65,536No LiveBench result for this model in release 2026-06-25322619.518.1

Each size is its own model, with its own scores and memory needs. Q4 estimates are planning figures, not measurements; file sizes are published. The Mac or PC column is for a 32 GB or larger machine (8 GB reserve); machines under 32 GB reserve 25% of their memory (2 to 8 GB), so they need less. The publisher lists 7B and 32B, each as Instruct and Think.

The two checkpoints listed here are different post-trained variants, 7B Instruct and 32B Think, so a difference between them mixes size and training. The family also has a 7B Think and a 32B Instruct (3.1) that are not in this site's catalog.

How to read this table

Q4 means compressed to about 4 bits per weight. GGUF (llama.cpp, Ollama, LM Studio) and MLX (Apple silicon) are file formats; the sizes shown are the published 4-bit files. A dash means no value is published or collected for that size; the reason is in the cell's accessible text.

Which size fits my machine

The largest size whose Q4 planning estimate fits what each machine offers. A planning estimate, not a measurement and not a speed.

Largest Olmo 3 size per machine
MachineLargest Olmo 3 size with enough memory
Mac mini M2 · 8 GB unified8Olmo 3 7B Instruct8
Mac mini M4 · 16 GB unified16Olmo 3 7B Instruct10
Mac mini M4 · 24 GB unified24Olmo 3 7B Instruct12
Mac mini M4 · 32 GB unified32Olmo 3 32B Think32
Mac mini M5 Pro · 64 GB unified56Olmo 3 32B Think32
Mac Studio M5 Max · 128 GB unified104Olmo 3 32B Think32
Mac Studio · M5 Ultra · 36-core CPU / 80-core GPU · 512 GB unified392Olmo 3 32B Think32
RTX 4060 · 8 GB VRAM8Olmo 3 7B Instruct8
RTX 3060 · 12 GB VRAM12Olmo 3 7B Instruct8
RTX 5080 · 16 GB VRAM16Olmo 3 7B Instruct8
RTX 4090 · 24 GB VRAM24Olmo 3 7B Instruct8
RTX 5090 · 32 GB VRAM32Olmo 3 32B Think26
DGX H100 · 8 × 80 GB640Olmo 3 32B Think40

A dash means no size of this family fits that machine. The requirement is the size's Q4 estimate with that machine's own reserve: a graphics card reserves 2 GB, as in the sizes table; a Mac or PC reserves 25% of its memory (2 to 8 GB), so below 32 GB its figure is lower than the sizes table's. A server profile reserves 2 GB per graphics card, so its figure is higher.

Smallest machine for each size

The profile with the least memory that still has enough for the size's Q4 estimate, from the hardware catalog.

Smallest machine per Olmo 3 size
SizeSmallest Mac or PC-memory profileSmallest graphics-card profile
Olmo 3 7B InstructMac mini 2020 · M1 · 8 GBRTX 3050 · 8 GB
Olmo 3 32B ThinkMac mini 2024 · M4 · 32 GBRTX 5090 · 32 GB

A dash means no profile of that kind in the catalog has enough memory for the size.

Memory by file precision

The same planning formula at three precisions (0.625, 1.125 and 2 bytes per parameter, plus working room and the system reserve). Estimates, not measurements.

Olmo 3 memory estimate by precision
Size
Olmo 3 7B Instruct14818122519
Olmo 3 32B Think322652468579

Family tree

Family, sizes and, for each size, the two 4-bit files the tables use (GGUF Q4_K_M and MLX 4-bit) when one was collected. Other precisions are not drawn here.

Olmo 3 family: sizes and published filesSwipe sideways to see the files, or read the list below.
Olmo 3 family: sizes and published filesFamilySizesFilesOlmo 3Ai27B InstructGGUF Q4_K_M4.5 GBMLX 4-bit4.1 GB32B ThinkGGUF Q4_K_M19.5 GBMLX 4-bit18.1 GB

The same tree as a list:

  • Olmo 3 · Ai2
    • 7B Instruct
      • GGUF Q4_K_M · 4.5 GB · unsloth/Olmo-3-7B-Instruct-GGUF, read Oct 23, 2026
      • MLX 4-bit · 4.1 GB · mlx-community/Olmo-3-7B-Instruct-4bit, read Oct 23, 2026
    • 32B Think
      • GGUF Q4_K_M · 19.5 GB · unsloth/Olmo-3-32B-Think-GGUF, read Oct 23, 2026
      • MLX 4-bit · 18.1 GB · lmstudio-community/Olmo-3-32B-Think-MLX-4bit, read Oct 23, 2026

Learn more

Sources

Where the family and its sizes are named, and where each number comes from.