List your company

Multi-task Evaluation on Spatial and Temporal Simulation Tasks Step-level

8.318Average Total ScoreCortex (full harness)
6.1866.73957.2937.8465Jul 6, 2026

Evaluation results

MethodLinks
Cortex (full harness)Level=Step-level2026.078.318
Cortex (harness on skills)Level=Step-level2026.077.392
Cortex (w/o harness)Level=Step-level2026.077.213
Cortex (baseline)Level=Step-level2026.077.051
GeminiLevel=Step-level2026.076.925
Qwen3-VL-8B-InstructLevel=Step-level2026.076.739
GPT-5Level=Step-level2026.076.268