A tabular foundation model predicts the missing column of any table, zero-shot, the way a language model completes text. On the main community benchmark, every single-model entry above the best tuned gradient-boosted tree is now one of these. This post explains what they are, verifies the strongest one with unrestricted open weights on my own hardware, and maps where the trees still win.
There is a fairly new class of model for tables: pretrained transformers that predict on any spreadsheet zero-shot, loosely called “tabular LLMs”. On the TabArena leaderboard, every single-model entry above the best tuned, ensembled GBDT configuration is one of them; the only entries above it that aren’t are AutoGluon’s 4-hour ensemble pipelines. That inverts the default answer of the last decade, which was to fit a tuned gradient-boosted tree. Leaderboards have been wrong before, in ways that only show up when an outsider re-runs them (that was my experience with time-series models), so before believing this one I audited it. I recomputed the board’s scoring from its published artefacts, took TabICLv2, the strongest model on it with unrestricted open weights, and re-ran it from scratch on hardware I controlled.
My independent run covered all 51 datasets on the benchmark’s official Lite protocol (one train/test split per dataset, the same split indices the leaderboard uses) on a single AWS A10G. Scored against the official artefact rows on those identical splits, my TabICLv2 lands at Elo 1559 against the official 1575, adjacent ranks and well inside the ±60–86 bootstrap intervals. Per task, 16 of the 51 metric values came back identical to four decimal places, the median relative difference was 0.08%, and the worst dataset differed by 3.5%, the size of gap you’d expect from GPU nondeterminism rather than a methodology difference. The whole sweep, environment setup included, fitted in a 2.1-hour GPU session that cost just over $2.
To continue reading this article, click here.
