From a Facebook post by Ted Underwood:
OLMo-2 from Allen AI is the first fully open GPT4-class model. Open code, open data, everything.
I think this is the door through which we get to a future where free intellectual work in the public interest is still possible. If we *couldn’t* pull this off, every path forward would lead to corporate enclosure of the intellectual commons.
A lot of people across different institutions and countries seem to be grasping the gravity of this challenge; there are efforts underway to get foundations, academic libraries, &c coordinated around this. Would love to see it happening even faster, but I guess the thing about freedom is people do what they want.
Ted's right about this, it would be a disaster if all the large LLMs, the so-called Foundation Models, were in private hands. Underwood's post links to Nathan Lambert, Gemma 3, OLMo 2 32B, and the growing potential of open-source AI, Mar. 13, 2025.
Ever since the release of the original ChatGPT, much has been said about making a truly open-source version of it — with data, code, weights, etc., all available. Open-source versions increase transparency, access, long-term progress, security research, and lots more. Lots of people have used this claim to bring hype into their projects, but the substance of these releases have been rather shallow (i.e., often focusing on one evaluation). [...]
Truly open releases take a lot of effort by making more to release and maintain, open up potential legal risks that preclude types of training data2, and completely undermine competition. The few organizations doing fully open-source research are non-profits, like Ai2 or Eleuther AI; academics, like LLM360; or companies that benefit from the long-term ecosystem growth, like HuggingFace.
I was poking through the results for our latest model when I realized that we finally did it! We have a fully open-source GPT-4 class model, i.e., it is comparable with OpenAI's original release rather than the current version.
The concluding paragraph:
The new part of this release is a major milestone where any company can pick up our training stack and cook up exactly the model they need at nearly the GPT 4 level. Beating the latest GPT 3.5 and GPT 4o mini models feels like fair game for the claim. This capability will take time to diffuse, but it is a major moment in the arc of why we do what we do. Even without more progress on OLMo, which we obviously will continue this year, this will keep fundamental AI progress outside of the major AI labs going for multiple years. It’s an optimistic day for open-source.
