Best score per boardGrok 4.6Step 5 Preview
DeveloperSpaceXAIStepFun
Artificial Analysis Intelligence Index v4.344.343.7
Terminal-Bench 4.020.3—
Terminal-Bench 4.0 (AA)21.233.3
Code Arena | WebDev Overall1620.01570.0
Agent Arena · Overall (Ranking / Pareto)1.2—
Artificial Analysis Coding Agent Index v1.547.0—
DeepSWE v1.167.5—
Radar IQ (Codex)107.1—
Lowest real price / MTok$0.026$0.0036

Data: real-api-pricingData: codexradarSources and licenses