Research map / Large language models / Agents, code and math
On Path to Multimodal Historical Reasoning: HistBench and HistAgent
Jiahao Qiu, Fulian Xiao, Yimin Wang, Yuchen Mao, Yijia Chen, Xinzhe Juan, Siran Wang, Xuan Qi, Tongcheng Zhang, Zixin Yao, Jiacheng Guo, Yifu Lu and 13 more
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use HistAgent (history-specific tools and manager-specialist agents) instead of generic Open Deep Research agent | GPT-4o agent · HistBench | pass@2 accuracy: 25.12 -> 36.47 pass@2; 19.57 -> 28.50 pass@1 | Improved |
| wrap GPT-4o in an agent framework (ODR-smolagents) versus GPT-4o with web search | GPT-4o · HistBench | accuracy: 18.60 -> 19.57 | Improved |
| use HistAgent (GPT-4o) versus stronger reasoning model with search | o4-mini-high with web search · HistBench | pass@1 accuracy: 30.19 vs 28.50 | Hurt |
| use HistAgent versus ODR-smolagents on HLE history questions | GPT-4o agent · HLE History subset | pass@1: 17.86 -> 28.57 | Improved |
| use HistAgent versus ODR-smolagents on general GAIA tasks | agent with strong backbone · GAIA validation | pass@1: 55.15 -> 60.00 | Improved |
| HistAgent versus ODR-smolagents on GAIA Level 3 | agent · GAIA validation Level 3 | pass@1: 34.62 -> 34.62 | No clear effect |
| use HistAgent on multimodal questions versus text-only | HistAgent (GPT-4o) · HistBench Level 1 modality split | pass@1: multimodal 35.63 vs ODR 11.49; text-only 21.52 vs ODR 21.52 | Mixed |
| allow second attempt (pass@2 versus pass@1) | HistAgent (GPT-4o) · HistBench | accuracy: 28.50 -> 36.47 | Improved |
Closest papers by meaning
- Large Language Models' Expert-level Global History Knowledge Benchmark (HiST-LLM) (NeurIPS 2024 · similarity 0.91)
- HSSBench: Benchmarking Humanities and Social Sciences Ability for Multimodal Large Language Models (ICLR 2026 · similarity 0.85)
- History Matters: Temporal Knowledge Editing in Large Language Model (AAAI 2024 · similarity 0.84)
- ChroKnowledge: Unveiling Chronological Knowledge of Language Models in Multiple Domains (ICLR 2025 · similarity 0.84)
- MME-SCI: A Comprehensive and Challenging Science Benchmark for Multimodal Large Language Models (AAAI 2026 · similarity 0.82)
- WenMind: A Comprehensive Benchmark for Evaluating Large Language Models in Chinese Classical Literature and Language Arts (NeurIPS 2024 · similarity 0.82)
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.
