Assistant Tasks on GAIA (official)
78.2Exact MatchTendem’s AI agent
Evaluation Results
| Method | Links | |
|---|---|---|
| Tendem’s AI agenthuman involvement=none2026.02 | 78.2 | |
| Flowith2026.02 | 78.1 | |
| Manus2026.02 | 73.4 | |
| HAL Generalist Agentbackbone=Claude Sonnet 4.52026.02 | 72.6 | |
| ChatGPT Deep Research2026.02 | 67.4 | |
| HF Open Deep Researchmodel=GPT-5 medium2026.02 | 62.8 |