Insight Generation on InsightBench 2025
69.3RecallDataSTORM
Evaluation results
| Method | Links | |
|---|---|---|
| DataSTORMBackbone LLM=gpt-5, LLM Judge=Qwen3-30B judge2026.04 | 69.3 | |
| DataSTORMBackbone LLM=gpt-5, LLM Judge=GPT-4o judge2026.04 | 61.9 | |
| AgentPoirotBackbone LLM=gpt-5, LLM Judge=Qwen3-30B judge2026.04 | 49.9 | |
| AgentPoirotBackbone LLM=gpt-5, LLM Judge=GPT-4o judge2026.04 | 47.1 |