LoHoSearch is a new benchmark of 544 KG-constructed questions across 11 domains where the strongest search agent scores 34.74% and context strategies add at most 6.8%.
DeepSearchQA: Bridging the comprehensiveness gap for deep research agents
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
years
2026 11representative citing papers
EvoBrowseComp is an auto-updatable benchmark of 800 complex questions for search agents, synthesized via a three-agent live-web framework to ensure temporal freshness and block parametric shortcuts.
SGR-Bench evaluates agentic LLM systems on state-gated retrieval tasks where evidence is only accessible after configuring site-specific states, with the strongest system reaching 66.18% item-level F1 and failures dominated by retrieval-scope drift.
DOLORES, an agent using a formal language for meta-reasoning to construct adaptive scaffolds on the fly, outperforms prior scaffolding methods by 24.8% on average across four hard benchmarks and multiple model sizes.
An interactive AI workbench for mathematicians achieves 48% on FrontierMath Tier 4 and helped solve open problems in early tests.
BRIGHT-Pro and RTriever-Synth advance reasoning-intensive retrieval by adding multi-aspect evidence evaluation and aspect-decomposed synthetic training, with the fine-tuned RTriever-4B showing gains over its base model.
ICBCBench is a new consortium-built benchmark that jointly measures retrieval-reasoning accuracy and end-to-end report quality for deep research agents in finance.
A 4B deep research agent trained on 10K open data outperforms prior agents under 9B parameters and narrows the gap to 30B-class systems on research benchmarks.
A tuned Bioptic Agent achieves 79.7% F1 on a new multilingual benchmark for global drug asset scouting, outperforming Gemini, Claude, GPT, and other models.
QuestBench is a student-constructed benchmark of 256 questions on which current deep research AI systems achieve a mean pass rate of 16.85% and a best-case rate of 57.58%.
citing papers explorer
-
LoHoSearch: Benchmarking Long-Horizon Search Agents Beyond the Human Difficulty Ceiling
LoHoSearch is a new benchmark of 544 KG-constructed questions across 11 domains where the strongest search agent scores 34.74% and context strategies add at most 6.8%.
-
EvoBrowseComp: Benchmarking Search Agents on Evolving Knowledge
EvoBrowseComp is an auto-updatable benchmark of 800 complex questions for search agents, synthesized via a three-agent live-web framework to ensure temporal freshness and block parametric shortcuts.
-
SGR-Bench: Benchmarking Search Agents on State-Gated Retrieval
SGR-Bench evaluates agentic LLM systems on state-gated retrieval tasks where evidence is only accessible after configuring site-specific states, with the strongest system reaching 66.18% item-level F1 and failures dominated by retrieval-scope drift.
-
Deep Reasoning in General Purpose Agents via Structured Meta-Cognition
DOLORES, an agent using a formal language for meta-reasoning to construct adaptive scaffolds on the fly, outperforms prior scaffolding methods by 24.8% on average across four hard benchmarks and multiple model sizes.
-
AI co-mathematician: Accelerating mathematicians with agentic AI
An interactive AI workbench for mathematicians achieves 48% on FrontierMath Tier 4 and helped solve open problems in early tests.
-
Rethinking Reasoning-Intensive Retrieval: Evaluating and Advancing Retrievers in Agentic Search Systems
BRIGHT-Pro and RTriever-Synth advance reasoning-intensive retrieval by adding multi-aspect evidence evaluation and aspect-decomposed synthetic training, with the fine-tuned RTriever-4B showing gains over its base model.
-
ICBCBench: An Industry Consortium Benchmark for Financial Deep Research
ICBCBench is a new consortium-built benchmark that jointly measures retrieval-reasoning accuracy and end-to-end report quality for deep research agents in finance.
-
DR-Venus: Towards Frontier Edge-Scale Deep Research Agents with Only 10K Open Data
A 4B deep research agent trained on 10K open data outperforms prior agents under 9B parameters and narrows the gap to 30B-class systems on research benchmarks.
-
Hunt Globally: Wide Search AI Agents for Drug Asset Scouting in Investing, Business Development, and Competitive Intelligence
A tuned Bioptic Agent achieves 79.7% F1 on a new multilingual benchmark for global drug asset scouting, outperforming Gemini, Claude, GPT, and other models.
-
Teaching AI Through Benchmark Construction: QuestBench as a Course-Based Practice for Accountable Knowledge Work
QuestBench is a student-constructed benchmark of 256 questions on which current deep research AI systems achieve a mean pass rate of 16.85% and a best-case rate of 57.58%.
- VibeSearchBench: Benchmarking Long-horizon Proactive Search in the Wild