Olmo 3
Team Olmo: Allyson Ettinger, Amanda Bertsch, Bailey Kuehl, David Graham, David Heineman, Dirk Groeneveld, Faeze Brahman, Finbarr Timbers, Hamish Ivison, Jacob Morrison, Jake Poznanski, Kyle Lo, Luca Soldaini, Matt Jordan, Mayee Chen, Michael Noukhovitch, Nathan Lambert
TL;DR
Olmo 3 introduces a family of fully open, large-scale language models designed for advanced reasoning, coding, and chat, with comprehensive details of their development process included.
Contribution
It presents the first complete lifecycle release of fully open models at 7B and 32B scales, emphasizing long-context reasoning and instruction following.
Findings
Olmo 3 Think 32B is the strongest fully-open thinking model to date.
The release includes all stages, checkpoints, and data used in model development.
Abstract
We introduce Olmo 3, a family of state-of-the-art, fully-open language models at the 7B and 32B parameter scales. Olmo 3 model construction targets long-context reasoning, function calling, coding, instruction following, general chat, and knowledge recall. This release includes the entire model flow, i.e., the full lifecycle of the family of models, including every stage, checkpoint, data point, and dependency used to build it. Our flagship model, Olmo 3 Think 32B, is the strongest fully-open thinking model released to-date.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
- 🤗allenai/Olmo-3.1-32B-Thinkmodel· 11k dl· ♡ 9811k dl♡ 98
- 🤗allenai/Olmo-3-7B-Thinkmodel· 208k dl· ♡ 95208k dl♡ 95
- 🤗allenai/Olmo-3-7B-Think-SFTmodel· 20k dl· ♡ 1020k dl♡ 10
- 🤗allenai/Olmo-3-32B-Think-SFTmodel· 1.0k dl· ♡ 41.0k dl♡ 4
- 🤗allenai/Olmo-3-32B-Think-DPOmodel· 1.3k dl· ♡ 41.3k dl♡ 4
- 🤗allenai/Olmo-3-7B-RL-Zero-Generalmodel· 190 dl· ♡ 8190 dl♡ 8
- 🤗allenai/Olmo-3-7B-RL-Zero-IFmodel· 238 dl· ♡ 7238 dl♡ 7
- 🤗allenai/Olmo-3-7B-RL-Zero-Mathmodel· 975 dl· ♡ 10975 dl♡ 10
- 🤗allenai/Olmo-3-7B-RL-Zero-Codemodel· 357 dl· ♡ 18357 dl♡ 18
- 🤗allenai/Olmo-3-7B-Instruct-SFTmodel· 28k dl· ♡ 428k dl♡ 4
- allenai/dolma3_mix-6Tdataset· 62k dl62k dl
- allenai/dolma3_mix-6T-1025-7Bdataset· 14k dl14k dl
- allenai/Dolci-Instruct-SFTdataset· 37k dl37k dl
- allenai/dolma3_dolmino_pooldataset· 8.2k dl8.2k dl
- allenai/dolma3_longmino_pooldataset· 22k dl22k dl
- allenai/dolma3_mix-150B-1025dataset· 47k dl47k dl
- allenai/dolma3_dolmino_mix-100B-1025dataset· 28k dl28k dl
- allenai/dolma3_dolmino_mix-10B-1025dataset· 4.4k dl4.4k dl
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.