AI Token Coach Burn Radar

A desk for leaks, mismatches and good practice across the AI capacity you already pay for. Every item is verified by a person before it appears here, and dated so you can see when it was last checked.

PillarAll Burn Radar Model Efficiency Context Craft Agent and Harness Efficiency Router Lab Workflow Clinic Team Practice Academy

ImpactAny High Medium Low Watch

BurnAny Up Down Neutral Unknown

2 itemsClear

Fri 11 SeptVerified 07:45
Agent and Harness Efficiency
Three Codex releases in a fortnight all point at context budgets
Subagent budget accounting, three-tier context management, and the end of summary compaction. Read together, the harness is moving from hiding context cost to exposing it.
MediumBurn downCurrent
Fri 11 SeptVerified 07:30
Router Lab
Frontier pricing converged, so the real choice moved down the stack
GPT-6 Astra and Claude Fable 5.1 now cost the same per token. With the top tier price-matched, the decision that moves your burn is which work never reaches the top tier at all.
HighBurn downCurrent

Three Codex releases in a fortnight all point at context budgets

Frontier pricing converged, so the real choice moved down the stack