AI Paper Atlas

Research map / Large language models / Agents, code and math

SealQA: Raising the Bar for Reasoning in Search-Augmented Language Models

Thinh Pham, Nguyen Phan Nguyen, Pratibha Zunjare, Weiyuan Chen, Yu-Min Tseng, Tu Vu

ICLR 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 64 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add FreshPrompt search (ablation)DeepSeek-R1-671B · Seal-Hardaccuracy (%): 22.4 -> 11.0Hurt
add FreshPrompt search (ablation)DeepSeek-R1-671B · Seal-Hard, never-changing questionsaccuracy (%): -17.7%Hurt
add FreshPrompt search (ablation)DeepSeek-R1-671B · Seal-Hard, 2025 questionsaccuracy (%): +5.4%Improved
add ChatGPT built-in searchGPT-4.1 · Seal-Hardaccuracy (%): +5.5%Improved
add FreshPrompt search instead of built-in searchGPT-4.1 · Seal-Hardaccuracy (%): 15.0 -> 14.6No clear effect
add built-in searchGPT-4.1-mini · Seal-Hardaccuracy (%): 13.8 -> 11.8Hurt
add FreshPrompt searchQwen3-235B-A22B · Seal-Hardaccuracy (%): +7.1%Improved
add FreshPrompt searchDeepSeek-R1-Distill-Qwen-14B · Seal-Hardaccuracy (%): +9.7%Improved
increase reasoning effort from low to high (ablation)o3 · Seal-0accuracy (%): 11.7 -> 17.1 -> 14.4Mixed
increase reasoning effort from low to high (ablation)o4-mini · Seal-0accuracy (%): 6.3 -> 5.4 -> 4.5Hurt
increase reasoning effort from low to high (ablation)o3-mini · Seal-0accuracy (%): 1.8 -> 2.7 -> 1.8No clear effect
provide only the gold document (no hard negatives) (ablation)GPT-4.1 · LongSealaccuracy (%): 48.0 (vs 15.0 w/o search)Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy