Post

Olmo 3 is notable as a "fully open" LLM - all of the training data is published, plus complete details on how the training process was run. I tried out the 32B thinking model and the 7B instruct models, + thoughts on why transparent training data is so important simonwillison.net/2025/Nov/22/...

Would it be possible to determine whether any of their data is from a questionable source, or would the scale make that impossible?

It's possible to directly search their training data, but it's pretty huge and you would have to download it and run searches against it yourself They use a crawl of the web as part of their training data, it's up to you if that counts as a "questionable source"