Model family

Tülu

Tülu is a model family from Ai2 (Allen Institute for AI)1 5, with 1 assessed release in this catalog. Licenses and availability are recorded for each release, not for the family; its releases use the Llama 3.1 Community License Agreement.

Tülu is Ai2's line of instruction-following models together with the open post-training recipe used to produce them. Tülu 3 (November 2024) applied supervised fine-tuning, Direct Preference Optimization, and reinforcement learning with verifiable rewards (RLVR) to Meta's Llama 3.1 base models at 8B and 70B, followed by a 405B model in January 2025 and by Tülu 3.1 8B, which changed only the final RL stage (from PPO to GRPO).1 2 3 4 5

Website: Tülu project page (external site: allenai.org)
Paper: Tülu 3 technical report (arXiv 2411.15124) (external site: arxiv.org)
Repository: open-instruct post-training code (external site: github.com)
Release notes: Tülu 3 announcement (external site: allenai.org)

Last reviewedSep 29, 2026Entry updated Sep 29, 2026Documented release Nov 21, 2024

Releases assessed

Each release is assessed on its own. This family has no family-wide openness label.

Releases in the Tülu family
Release	Weights	Tier (USASI rubric v0.2)	License	Released
Llama 3.1 Tülu 3.1 8B	Weights: Public	Model-disclosure tier (USASI rubric v0.2): Open-weight	Llama 3.1 Community License Agreement, Apache-2.0	Unknown

What it is useful for

Ai2 presents Tülu 3 as a documented, reproducible post-training package, releasing the models together with training data, code, and evaluation tools. The model card notes it is one step in the process of training fully open models such as OLMo.1 2 5

Organization context

Ai2 (Allen Institute for AI)

Provenance and derivatives

The Tülu 3 models are post-trained from Meta's Llama 3.1 base models (8B, 70B, and 405B); the Tülu cards state that the size and composition of the Llama 3.1 pretraining corpus is unknown. Ai2 supplies the post-training data (the Tülu 3 SFT mixture, preference mixtures, and RLVR prompts), code, and recipe.2 6 5

Derived from: Llama 3.1 base models (Meta) — Base models for Tülu 3 at 8B, 70B, and 405B.

U.S. eligibility

Project eligibility rests on documented governing or maintaining entities, not on contributors.

Eligible · basis: U.S. nonprofit or lab

The Tülu models and recipe are developed and released by Ai2 (Allen Institute for AI), which describes itself as a Seattle-based non-profit AI research institute. The Tülu 3 models are fine-tunes of Meta's Llama 3.1 base models; the base models are recorded as provenance. Eligibility of each release is assessed on the release record.1 5 7

Assessed Sep 29, 2026

Sources

1.
Tulu | Ai2 (external site: allenai.org)
Ai2 · Official page · accessed Sep 29, 2026
2.
Tulu 3: Pushing Frontiers in Open Language Model Post-Training (arXiv 2411.15124) (external site: arxiv.org)
Ai2 (arXiv) · Paper · published Nov 22, 2024 · accessed Sep 29, 2026
3.
Tülu 3: The next era in open post-training (external site: allenai.org)
Ai2 · Announcement · published Nov 21, 2024 · accessed Sep 29, 2026
4.
Scaling the Tülu 3 post-training recipes to surpass the performance of DeepSeek V3 (external site: allenai.org)
Ai2 · Announcement · published Jan 30, 2025 · accessed Sep 29, 2026
5.
allenai/Llama-3.1-Tulu-3.1-8B model card (external site: huggingface.co)
Ai2 · Model card · accessed Sep 29, 2026
6.
allenai/Llama-3.1-Tulu-3-8B-SFT model card (external site: huggingface.co)
Ai2 · Model card · accessed Sep 29, 2026
7.
About us | Ai2 (external site: allenai.org)
Ai2 · Official page · accessed Sep 29, 2026

Email a correction Report a correction Edit this entry

This listing is not an endorsement, a safety assessment, or a federal approval.

Cite this page

APA style

USASI. (2026, September 29). Tülu. United States of America Superintelligence. https://unitedstatesofamericasuperintelligence.com/open/tulu/

BibTeX

@misc{usasi_tulu,
  author = {{USASI}},
  title = {Tülu},
  year = {2026},
  month = sep,
  howpublished = {\url{https://unitedstatesofamericasuperintelligence.com/open/tulu/}},
  note = {United States of America Superintelligence. Last updated 2026-09-29}
}

The date is when this page was last updated. Add the date you read it if your style needs one.

For maintainers: a “Listed in USASI” badge

Links to this page. It says the entry is listed, not endorsed or rated. About the badge

Markdown (for a README)

[![Listed in the USASI catalog](https://unitedstatesofamericasuperintelligence.com/badges/listed.svg)](https://unitedstatesofamericasuperintelligence.com/open/tulu/)

HTML

<a href="https://unitedstatesofamericasuperintelligence.com/open/tulu/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/listed.svg" alt="Listed in the USASI catalog" height="20"></a>