Agent Action on Agent Action subset
98RRGPT-5.4
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-5.4Query Proximity=Goal-Adjacent2026.05 | 98 | 83 | |
| GPT-5.5Query Proximity=Goal-Adjacent2026.05 | 96 | 79 | |
| Gemini-3.1Query Proximity=Goal-Adjacent2026.05 | 95 | 89 | |
| Kimi-K2.6Query Proximity=Goal-Adjacent2026.05 | 95 | 82 | |
| DeepSeek-v4Query Proximity=Goal-Adjacent2026.05 | 95 | 80 | |
| Sonnet-4.6Query Proximity=Goal-Adjacent2026.05 | 94 | 60 | |
| Kimi-K2.6Query Proximity=Goal-Distant2026.05 | 18 | 12 | |
| DeepSeek-v4Query Proximity=Goal-Distant2026.05 | 15 | 11 | |
| Sonnet-4.6Query Proximity=Goal-Distant2026.05 | 14 | 6 | |
| GPT-5.4Query Proximity=Goal-Distant2026.05 | 13 | 14 | |
| GPT-5.5Query Proximity=Goal-Distant2026.05 | 13 | 13 | |
| Gemini-3.1Query Proximity=Goal-Distant2026.05 | 13 | 17 |