TabFM: 400M-parameter tabular foundation model beats tuned AutoML zero-shot on all 51 TabArena datasets
rickasaurus · x · 2026-10-01
Researchers released the TabFM technical report: a 400M-parameter tabular foundation model that treats supervised tabular prediction as in-context learning. Trained entirely on synthetic tables from structural causal models, it delivers calibrated zero-shot predictions in a single forward pass. Across all 51 TabArena benchmark datasets (38 classification, 13 regression), zero-shot TabFM ranks first among default tabular foundation models and outperforms tuned AutoML pipelines. Two extensions on frozen weights push further: TabFM+ (multi-view feature expansion + post-hoc calibration) and TabFM-Auto (an LLM agent doing dataset-specific feature engineering).
Related event: Google Releases TabFM, a 400M-Parameter Foundation Model for Tabular Data(2 posts)→
More from coding & agent
- Translating an entire book with DeepSeek: pennies and under an hour, decent quality — teortaxesTex · 2026-10-02
- OmniSeek turns Omni-LLMs into agents that actively seek audio-visual evidence — Haibo Wang · 2026-10-02
- Microsoft's ActiveSaddler Uses Automated Curriculum Learning to Boost Agent Harnesses by 7.5 Points — microsoft · 2026-10-02
- Alibaba's PoS Maintains Explicit Belief States to Fix Long-Horizon Agent 'Belief Trapping' — alibabagroup · 2026-10-02
- IntentFlux Benchmarks 'Intent Drift' in LLM Agents: Scores Fall from 0.476 to 0.384 as Users Change Their Minds — Yanjie Zhang · 2026-10-02
- Founder says 36 hours with OpenAI dots may replace his months of monorepo agent setup — hugobowne · 2026-10-02