Grok usage meter bug hides sub-agent costs, underbilling by 4x
Daniel_Farinax · x · 2026-09-02
A user discovered a critical bug in Grok's usage meter. During a 5-hour session, the dashboard showed a consumption of 17% quota ($32). However, the loop spawned 45 sub-agent sessions that were not counted by the /usage command, adding another $110 in hidden costs. The real total was $143 (547M tokens), roughly 4.4x the displayed amount. In comparison, Claude's tracking was accurate.
Related event: Grok Heavy billing underreports hidden sub-sessions(2 posts)→
More from Models
- User Test: Fable 5.1 Outperforms GPT 5.6 on Complex Profiling Task — remilouf · 2026-09-02
- RTX 5090 laptop owner asks: best abliterated Qwen 3.6 vs 3.8 27B quants for 24GB VRAM — ThomasAger · 2026-09-02
- Merge Gateway Adds Claude Fable 5.1 with 75% Cheaper Cache Reads — shensi · 2026-09-02
- Daily AI brief: OpenAI's cyber-critical Astra, Grok 4.7 dated, Fable 5.1 live — testingcatalog · 2026-09-02
- OpenAI's 'Astra' rumored to launch this week with new recurrent depth architecture — mark_k · 2026-09-02
- iFlytek's Spark-X2.5-4B trends on Hugging Face — XHToken · 2026-09-02