Open the live control panel | Research manual

Model and effort allocation

Codex arm defaults. OpenRouter reader arm and configuration. Pilot allocation, not a measured optimum.

StageEffort by stepReaders
A1a: medium, b: medium, c: highnone
B1a: high, b: high, c: xhigh, d: high, e: xhighnone
B2a: medium, b: high, c: medium, d: highgpt-6-sol medium
B3a: medium, b: highgpt-6-sol medium
B4a: high, b: medium, c: highgpt-6-sol medium
C1a: xhighnone
C2a: xhigh, b: high, c: high, d: xhighgpt-6-astra high
C3a: high, b: high, c: xhighnone
C4a: highnone
C5a: high, b: medium, c: high, d: highgpt-6-astra high
D1a: mediumnone
D2a: highgpt-6-sol medium
D3a: high, b: medium, c: highnone
D4a: xhighnone
D5a: high, b: highnone
D6a: high, b: highnone
E1a: high, b: high, c: medium, d: high, e: highgpt-6-sol medium
E2a: high, b: xhighnone