Copilot's reasoning_tokens are parsed but not added to output tokens; OpenCode's are

Copilot’s ingest payload includes a reasoning_tokens field per prompt, but it is intentionally not added into the stored output-token count. OpenCode’s equivalent field is added to its output-token count.

This asymmetry is correct: Copilot already includes reasoning tokens in its reported output-token total, so adding them again would double-count. OpenCode reports reasoning tokens separately from its output-token total, so they must be added explicitly to get the true output-token count. The reasoning_tokens field on the Copilot side is parsed because it is present in every payload, and then deliberately left unused.