Skip to content
AI Atlas
Familyroot · OLMo
data quality35

first seen 11 Sept 2026

family_01M2A55V4FJBBB542RKFGNRVHC

Models
5
First release
18 Nov 2025
Latest release
19 Nov 2025
Parameters
7.3B
Artifacts
0
Providers
0

Releases

Family releases by statusActiveNov 25Nov 25Nov 25Nov 25Olmo-3-7B-Think — 18 Nov 2025Olmo-3-7B-Instruct — 19 Nov 2025

2 dated releases · dot size = best benchmark rank of the member · 3 members without a sourced release date not shown

Members 5

Family members
ModelParamsContextOpennessLicenceReleasedStatusBest rankCompare
Olmo-3-7B-InstructAllen Institute for AI7.3B65.5KOpen weightsApache-2.019 Nov 2025Active#258 IFBench
Olmo-3-7B-ThinkAllen Institute for AI7.3B65.5KOpen weightsApache-2.018 Nov 2025Active#190 IFBench
olmo-3-32b-thinkAllen Institute for AI65.5KOpen weightsDeprecated#153 IFBench
Olmo-3-7B-Instruct-DPOAllen Institute for AIOpen weightsActive
Olmo-3-7B-Think-DPOAllen Institute for AIOpen weightsActive

Lineage among members 2 relations

Open in Graph →
Lineage of the OLMo 3 familyOlmo-3-7B-Instruct fine-tuned from Olmo-3-7B-Instruct-DPOOlmo-3-7B-Think fine-tuned from Olmo-3-7B-Think-DPOOlmo-3-7B-Think7.3BOlmo-3-7B-ThinkOlmo-3-7B-Instruct7.3BOlmo-3-7B-InstructOlmo-3-7B-Instruct-DPOOlmo-3-7B-Instruct-DPOOlmo-3-7B-Think-DPOOlmo-3-7B-Think-DPO

Benchmark progress

IFBenchbest #190

Best rank of the family on IFBench#250#200#150#100#50#0Nov 25Nov 25Nov 25Nov 25

Terminal-Benchbest #265

Best rank of the family on Terminal-Bench#400#300#200#100#0Nov 25Nov 25Nov 25Nov 25

Humanity's Last Exambest #273

Best rank of the family on Humanity's Last Exam#400#300#200#100#0Nov 25Nov 25Nov 25Nov 25

τ²-benchbest #282

Best rank of the family on τ²-bench#400#300#200#100#0Nov 25Nov 25Nov 25Nov 25

GPQA Diamondbest #317

Best rank of the family on GPQA Diamond#400#300#200#100#0Nov 25Nov 25Nov 25Nov 25

Artificial Analysis Intelligence Indexbest #408

Best rank of the family on Artificial Analysis Intelligence Index#500#400#300#200#100#0Nov 25Nov 25Nov 25Nov 25

Best rank (primary comparability group, current leaderboard) reached by any member released up to each date — lower is better, axis inverted. Ranks are today's ranks, not the ranks at release time.

Aggregates are computed from the members' sourced claims; members without a release date or parameter count are left out of the range and timeline rather than guessed. Methodology →