GAIA

GAIA dataset GAIA is a benchmark which aims at evaluating next-generation LLMs (LLMs with augmented capabilities due to added tooling, efficient prompting, access to search, etc). We added gating to prevent bots from scraping the dataset. Please do not reshare the validation or test set in a crawlable format. Data and

Papers using GAIA (12)

Beyond Reward Engineering: A Data Recipe for Long-Context Reinforcement Learning2026Beyond Ten Turns: Unlocking Long-Horizon Agentic Search with Large-Scale Asynchronous RL2025 · 1 citesSlimSearcher: Training Efficiency-Aware Web Agents via Adaptive Reward Gating2026Maximizing Rollout Informativeness under a Fixed Budget: A Submodular View of Tree Search for Tool-Use Agentic Reinforcement Learning2026EvoMAS: Learning Execution-Time Workflows for Multi-Agent Systems2026LiteResearcher: A Scalable Agentic RL Training Framework for Deep Research Agent2026Exploring Reasoning Reward Model for Agents2026Youtu-Agent: Scaling Agent Productivity with Automated Generation and Hybrid Policy Optimization2025Agentic Entropy-Balanced Policy Optimization2025DeepAgent: A General Reasoning Agent with Scalable Toolsets2025Tool-R1: Sample-Efficient Reinforcement Learning for Agentic Tool Use2025Memento: Fine-tuning LLM Agents without Fine-tuning LLMs2025