2026
SealQA: Raising the Bar for Reasoning in Search-Augmented Language Models
ICLR 2026poster
We introduce SealQA, a challenge benchmark for evaluating SEarch-Augmented Language models on fact-seeking questions where web search yields conflicting, noisy, or unhelpful results. SealQA comes in three flavors: (1) SEAL-0 (main) and (2) SEAL-HARD, both of which assess factual accuracy and reasoni…