Two Models Show Massive Divergence in Paper Reproduction

mgostIH · x · 2026-07-16

The author tasked **Sol Ultra** and **Fable Max** with reproducing a paper's conclusions, resulting in a stark divergence: - **Sol Ultra**: Concluded the findings were "untenable" and the reproduction failed. - **Fable Max**: Concluded the findings were "completely valid." Though the post has a slightly sarcastic tone, the core takeaway is that different models can have vastly different judgments on the same research reproduction task, leaving the author to manually verify the results.

Related event: Models Split on Reproducing a Paper’s Claim(2 posts)→

Original post →

More from Models

Models channel →