Multi-task Evaluation on Spatial and Temporal Simulation Tasks Episode-level
7.81Average Total ScoreCortex (full harness)
Evaluation Results
| Method | Links | |
|---|---|---|
| Cortex (full harness)Level=Episode-level2026.07 | 7.81 | |
| Cortex (harness on skills)Level=Episode-level2026.07 | 7.314 | |
| GPT-5Level=Episode-level2026.07 | 7.231 | |
| Cortex (w/o harness)Level=Episode-level2026.07 | 7.155 | |
| Cortex (baseline)Level=Episode-level2026.07 | 6.978 | |
| GeminiLevel=Episode-level2026.07 | 6.86 | |
| Qwen3-VL-8B-InstructLevel=Episode-level2026.07 | 6.292 |