# TabPFN

10 Hacker News commenters recommended TabPFN and 0 criticized it, November 2025 to November 2026. It was recommended in 10 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Recommended for accurate tabular prediction with no tuning, since one forward pass matches tuned AutoGluon and it handles messy real-world data. Main caveat: inference is slower than traditional models.

## Why people recommend it

- State of the art on tabular benchmarks (3 comments): "TabPFN was already state of the art and is seriously shaking up Bayesian prediction for tabular data" ([source](https://news.ycombinator.com/item?id=48740233)) "v2 was published in Nature and set a new state of the art on tabular benchmarks" ([source](https://news.ycombinator.com/item?id=48360453)) "TabPFN v2 was published in Nature and set a new state of the art" ([source](https://news.ycombinator.com/item?id=49165353))
- Foundation model that understands tabular data (2 comments): "Tabular foundation models like TabPFN and related work are extremely promising." ([source](https://news.ycombinator.com/item?id=47257033)) "the first foundation model that actually understands tabular data" ([source](https://news.ycombinator.com/item?id=47219823))
- Big gains in version 2.5 (2 comments): "TabPFN-2.5 default (one forward pass) matches AutoGluon 1.4 tuned for four-hours" ([source](https://news.ycombinator.com/item?id=45839362)) "the improvements with 2.5 look massive" ([source](https://news.ycombinator.com/item?id=45891248))

## What people warn about

- Slow inference compared with traditional models (1 comments): "The inference time is still much worse than any of the more traditional models" ([source](https://news.ycombinator.com/item?id=49171827))

## Cite as

Hacker Approved, "TabPFN", https://hackerapproved.com/p/tabpfn/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).