2 replies from Opus 4.6 consumed 90% of usage meter

Not sure if this is a bug, but usage limit suddenly became too low lately?

I’ll attach the exported thread, since it’s not transparent how much usage per token is alloted there really is, only t3.chat team can analyze if the consumption is the meant 5-hour limit (but feels too restrictive for a pro plan if it is compared to the past)

tried to ask glm-5 to analyze the thread, here’s its analysis
```

Analysis Summary JSON Structure The export file has this structure: { "threads": [...] // Thread metadata (title, model, timestamps) "messages": [...] // Actual messages with token usage in providerMetadata.anthropic.usage "version": "..." } Token usage is nested in each message at: message.providerMetadata.anthropic.usage Your Thread's Token Consumption

```