Module FTDD-02 — OLMo 2/3 + Tülu 3 (Ai2)

OLMo 2/3 + Tülu 3 (Ai2)

The fully-open comparison: Allen Institute's OLMo (the open base) and Tülu 3 (the open SFT→DPO→RLVR post-training recipe). The research-oriented counterpoint to MiniCPM's product orientation.

45
minutes
8
artifacts
5
sub-sections
What 'open' means when a lab commits to releasing everything. OLMo 2 (arXiv:2501.00656) established the fully-open base standard; OLMo 3 (Nov 2025, base/instruct/think) makes each steering stage comparable on one base; Tülu 3 (arXiv:2411.15124) releases the full post-training pipeline (SFT → DPO → RLVR) up to 405B. The strict end of the open spectrum — rebuildable, ablatable, auditable end-to-end.
Key Claims
Load-Bearing Claims

OLMo and Tülu together cover the full open stack: the base AND the post-training recipe. OLMo (arXiv:2501.00656) is the open base — weights, data, code, checkpoints, eval. Tülu 3 (arXiv:2411.15124) is the open post-training pipeline — SFT, DPO, RLVR, each released as separable components. Both needed for full reproducibility.

OLMo 3's three variants (base/instruct/think) are a pedagogical gift. The same base with three different steers (none / SFT+DPO / +RLVR), directly comparable. A student can load all three and SEE what each steering stage changed — the FT00 steering stack made concrete, the clearest illustration of FT12/FT13/FT14.

Tülu 3's released pipeline enables ablation — the property that weights-only releases cannot provide. Because each stage (SFT, DPO, RLVR) is released as data + code + configs, a researcher can train Tülu 3 with one stage removed and measure the effect. This is how research advances, and it is impossible when the data and recipe are opaque.

MiniCPM is what you ship; OLMo/Tülu is what you study. Both open-data/open-recipe (Apache-2.0), but different centers of gravity: MiniCPM's variant axis is modality (product/edge); OLMo 3's variant axis is the steer (research/audit). MiniCPM for small edge models; OLMo/Tülu for reproducing, auditing, and understanding the full stack.

After This Module
01
Distinguish Ai2's OLMo (the open base) from Tülu 3 (the open post-training recipe) and explain why both are needed for full reproducibility.
02
Map the OLMo 2 → OLMo 3 progression and the base/instruct/think variants, citing arXiv:2501.00656 (OLMo 2) and arXiv:2411.15124 (Tülu 3).
03
Trace the Tülu 3 post-training pipeline (SFT → DPO → RLVR) and place each stage on the FT00 steering stack.
04
Defend why fully-open (weights + data + code + recipe + checkpoints) matters for reproducibility, ablation, and supply-chain trust.
05
Contrast the OpenBMB/MiniCPM (product/edge) and Ai2/OLMo-Tülu (research/audit) philosophies and choose between them.
Artifacts