TabPFN
Hacker ApprovedTabular foundation model in AI tools
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.
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.
10 people gave an opinion
- recommend it
- 10
- criticize it
- 0
- approval share
- 100%
Recommended in 10 of 13 months
- November 2025: recommended
- December 2025: recommended
- January 2026: not recommended
- February 2026: not recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 10 people recommend itNeeds 10 or more, met.
- Nobody criticizes itNeeds 2 to 1 or better, met.
- Recommended in 10 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
State of the art on tabular benchmarks
“TabPFN was already state of the art and is seriously shaking up Bayesian prediction for tabular data”
Comment from 30 Jun 2026 in “TabFM: A zero-shot foundation model for tabular data” “v2 was published in Nature and set a new state of the art on tabular benchmarks”
Comment from 1 Jun 2026 in “Ask HN: Who is hiring? (June 2026)” “TabPFN v2 was published in Nature and set a new state of the art”
Comment from 4 Aug 2026 in “Ask HN: Who is hiring? (August 2026)” Foundation model that understands tabular data
“Tabular foundation models like TabPFN and related work are extremely promising.”
Comment from 5 Mar 2026 in “Tabular Foundation Models Still Need One Thing: Multi-Table Aggregation” “the first foundation model that actually understands tabular data”
Comment from 2 Mar 2026 in “Ask HN: Who is hiring? (March 2026)” Big gains in version 2.5
“TabPFN-2.5 default (one forward pass) matches AutoGluon 1.4 tuned for four-hours”
Comment from 6 Nov 2025 in “Show HN: TabPFN-2.5 – SOTA foundation model for tabular data” “the improvements with 2.5 look massive”
Comment from 11 Nov 2025 in “Show HN: TabPFN-2.5 – SOTA foundation model for tabular data”
What people warn about
Each mark is one comment
Slow inference compared with traditional models
“The inference time is still much worse than any of the more traditional models”
Comment from 4 Aug 2026 in “Why Large Language Models Fail at Tabular Prediction”
What people said
14 of the 19 comments that mention it, newest first. Each links to the original comment.
“But most importantly it really improves on the messy real-world data we've seen across industries.”
“The inference time is still much worse than any of the more traditional models”
“Non-LLM transformers beat tree boosting”
“TabPFN v2 was published in Nature and set a new state of the art”
“TabPFN was already state of the art and is seriously shaking up Bayesian prediction for tabular data”
“v2 was published in Nature and set a new state of the art on tabular benchmarks”
“Tabular foundation models like TabPFN and related work are extremely promising.”
“the first foundation model that actually understands tabular data”
“unless TabPFNv2 makes sense for your use-case”
“the improvements with 2.5 look massive”
“the API is the best way to handle text features”
“Glad its useful to people!”
“TabPFN-2.5 default (one forward pass) matches AutoGluon 1.4 tuned for four-hours”
“I think it's a really creative application of neural networks”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.