General AI Assistants Evaluation on GAIA text-103
70.9AccuracyTongyi-DR
Evaluation results
| Method | Links | |
|---|---|---|
| Tongyi-DRcrawl(s)=✓, visual=✗, file=✗, browser=✗, Backbone=TongyiDR-30B-A3B† + GPT-4o2026.03 | 70.9 | |
| MiroThinker v0.1crawl(s)=✓, visual=✓, file=✓, browser=✓, Backbone=MiroThinker-32B-DPO + GPT-4.1 +Claude-S3.72026.03 | 57.9 | |
| WebSailorcrawl(s)=✓, visual=✗, file=✗, browser=✗, Backbone=WebSailor-32B + Qwen3-72B2026.03 | 53.2 |