Conversational agent benchmark — simulated retail customer service tasks measuring tool-calling, order management, and multi-turn dialogue reliability.
Complements airline scenarios with retail tasks such as inventory lookups, returns, and order updates; evaluates whether agents preserve context across transactional multi-turn interactions.
Limited to retail domain. Small model roster in the current leaderboard.
Higher is better. Pass^1 rate.
Max score: 100
Limited to retail domain.
| # | Model | Score |
|---|---|---|
| 1 | 99.6 | |
| 2 | 99.3 | |
| 3 | 98.9 | |
| 4 | 95.8 | |
| 5 | 95.7 | |
| 6 | 95.7 | |
| 7 | 95.4 | |
| 8 | 95.3 | |
| 9 | 94.9 | |
| 10 | 94.9 | |
| 11 | 94.7 | |
| 12 | 94.3 | |
| 13 | 94.3 | |
| 14 | 94.0 | |
| 15 | 94.0 | |
| 16 | 93.6 | |
| 17 | 93.5 | |
| 18 | 93.4 | |
| 19 | 93.4 | |
| 20 | 93.2 | |
| 21 | 93.2 | |
| 22 | 93.1 | |
| 23 | 93.0 | |
| 24 | 92.8 | |
| 25 | 92.8 | |
| 26 | 92.4 | |
| 27 | 92.2 | |
| 28 | 92.2 | |
| 29 | 92.2 | |
| 30 | 92.2 | |
| 31 | 92.2 | |
| 32 | 92.2 | |
| 33 | 92.1 | |
| 34 | 92.1 | |
| 35 | 91.9 | |
| 36 | 91.9 | |
| 37 | 91.9 | |
| 38 | 91.7 | |
| 39 | 91.6 | |
| 40 | 91.6 | |
| 41 | 91.5 | |
| 42 | 91.4 | |
| 43 | 91.4 | |
| 44 | 91.4 | |
| 45 | 91.4 | |
| 46 | 91.4 | |
| 47 | 91.3 | |
| 48 | 91.2 | |
| 49 | 91.1 | |
| 50 | 91.0 |