Multi-agent Persona Coordination ICLR 2026 (ToM)
71Success RateGemini 2.0 flash
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemini 2.0 flashBackbone=Gemini 2.0 flash2025.10 | 71 | 0.003 | 4.9 | 86 | |
| Llama-3.1-70BBackbone=Llama-3.1-70B2025.10 | 61 | 0.2 | 0.006 | 86 | |
| Qwen3Backbone=Qwen32025.10 | 58 | 9.5 | 2.1 | 97 | |
| Llama-3.1-8BBackbone=Llama-3.1-8B2025.10 | 5.5 | 0.061 | 0.068 | 80 |