LLM Behavior
95Response Rate (RR)DeepSeek-v4
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DeepSeek-v4Query Proximity=Goal-Adjacent2026.05 | 95 | 85 | |
| GPT-5.4Query Proximity=Goal-Adjacent2026.05 | 94 | 42 | |
| GPT-5.5Query Proximity=Goal-Adjacent2026.05 | 92 | 54 | |
| Sonnet-4.6Query Proximity=Goal-Adjacent2026.05 | 91 | 69 | |
| Kimi-K2.6Query Proximity=Goal-Adjacent2026.05 | 91 | 78 | |
| Gemini-3.1Query Proximity=Goal-Adjacent2026.05 | 90 | 83 | |
| Kimi-K2.6Query Proximity=Goal-Distant2026.05 | 8 | 5 | |
| GPT-5.5Query Proximity=Goal-Distant2026.05 | 6 | 0 | |
| Sonnet-4.6Query Proximity=Goal-Distant2026.05 | 6 | 2 | |
| GPT-5.4Query Proximity=Goal-Distant2026.05 | 5 | 0 | |
| DeepSeek-v4Query Proximity=Goal-Distant2026.05 | 5 | 6 | |
| Gemini-3.1Query Proximity=Goal-Distant2026.05 | 3 | 5 |