OpenCodePapers

agents-on-gaia-text-only-103-validation-tasks

Agents
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeAccuracyModelNameReleaseDate
AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective Tasks✓ Link92.2AREX-22026-09-29
QUEST: Training Frontier Deep Research Agents with Fully Synthetic Tasks✓ Link80.8QUEST-35B2026-05-22
Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale77.4Intern-S1-Pro2026-03-26
Nanbeige4.1-3B: A Small General Model that Reasons, Aligns, and Acts✓ Link69.90Nanbeige4.1-3B (w/ tools, text-only)2026-02-13
DIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool Use✓ Link61.2Dive-8B (RL)2026-03-10
ANTMAN: Adaptive Need Tracking for Multi-Agent Navigation in Large Information Spaces58.3ANTMAN2026-09-27
ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RL✓ Link57.93ContextPilot2026-08-28
IterSynth: Rethinking Deep Search Agents via Role-Decoupled Iterative Synthesis✓ Link55.3IterSynth-8B2026-09-24
Learning Agentic Policy from Action Guidance✓ Link41.74ActGuide-RL (Qwen3-8B)2026-05-12
Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models✓ Link33.9Youtu-LLM 2B2025-12-31
Check out how to contribute new results. Then edit this file on GitLab or GitHub.