AI BriefingPrior Labs Releases TabPFN-3.5, a Foundation Model for Tabular Data
Key point
Prior Labs has released TabPFN-3.5, a foundation model that significantly improves performance in tabular data processing.
Details
Prior Labs has released its latest tabular data foundation model, TabPFN-3.5. The model achieved first place on both the TabArena and BeyondArena benchmarks, reaching State-of-the-Art (SOTA) performance on datasets with 1 million rows and up to 20,000 features.
Key Model Variants and Features
- TabPFN-3.5-Fast (Alpha Version): Offers inference speeds 6x faster than the base model.
- TabPFN-3.5-Thinking: Available via API, this variant increases computational effort to improve accuracy. It showed performance improvements of +20 Elo on BeyondArena and +44 Elo on TabArena compared to the base model.
- TabPFN-3.5-Plus: A version with additional performance enhancements.
Benchmark Performance On BeyondArena, TabPFN-3.5 excelled in handling data rich in text with high cardinality and dimensionality. It demonstrated a performance advantage with score differences of +250 Elo over the strongest previous baseline and +150 Elo over the previous overall top model.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Step 5 Preview Released: 600B Parameter MoE Model Highlights Agentic Tasks and Financial Analysis Performance