Allen AI/Dense

OLMo 3.1 32B Think

Name: OLMo 3.1 32B Think
Author: Allen AI

We introduce Olmo 3, a new family of 7B and 32B models both Instruct and Think variants. Long chain-of-thought thinking improves reasoning tasks like math and coding.

chatreasoningmathcoding

32B

Parameters

64K

Context length

Benchmarks

Quantizations

Architecture

Dense

Released

2026-01-20

Layers

KV Heads

Head Dim

128

Family

olmo

Quantization Options

Quant	Bits	VRAM	Quality
IQ3_XXS	3.25	13.5 GB	low
IQ3_XS	3.5	14.5 GB	low
Q3_K_S	3.64	15.0 GB	low
IQ3_M	3.76	15.5 GB	low
Q3_K_M	4	16.5 GB	low
Q3_K_L	4.3	17.7 GB	moderate
IQ4_XS	4.46	18.3 GB	moderate
Q4_K_S	4.67	19.2 GB	moderate
Q4_K_M	4.89	20.0 GB	good
Q5_K_S	5.57	22.8 GB	good
Q5_K_M	5.7	23.3 GB	good
Q6_K	6.56	26.7 GB	excellent
Q8_0	8.5	34.5 GB	lossless
FP16	16	64.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN OLMO 3.1 32B THINK NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (11)

AIME77.3

AA Math77.3

MATH-50077.3

MMLU-PRO76.3

LiveCodeBench69.5

IFBench66.0

GPQA Diamond59.1

SciCode29.3

AA Intelligence13.9

AA Coding9.8

HLE6.0

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run olmo:32b-q4_K_M

Tag may need adjustment — check ollama.com/library/olmo for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware

GPUs that can run this model

At Q4_K_M quantization. Sorted by minimum VRAM.

NVIDIA RTX 4090

24 GB VRAM • 1008 GB/s

NVIDIA

$1599