fix: stop deducting simulated cache tokens from input_tokens
Some checks failed
Build Docker Image / build (push) Has been cancelled

Kiro backend does not support Anthropic prompt cache protocol.
The local cache tracker simulates cache hits/creation for Claude Code
compatibility, but subtracting those values from input_tokens caused
the reported input_tokens to drop to single digits.

input_tokens now reflects the real value; cache_creation_input_tokens
and cache_read_input_tokens are still reported for protocol compliance.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-05-12 11:14:51 +08:00
parent c87517c0bf
commit e8ab5b11e7
3 changed files with 4 additions and 8 deletions

View File

@@ -1286,7 +1286,7 @@ func (h *Handler) handleClaudeNonStream(w http.ResponseWriter, account *config.A
}
resp := KiroToClaudeResponse(finalContent, responseThinkingContent, includeEmptyThinkingBlock, toolUses, inputTokens, outputTokens, model)
resp.Usage.InputTokens = billedClaudeInputTokens(inputTokens, cacheUsage)
resp.Usage.InputTokens = inputTokens
resp.Usage.CacheCreationInputTokens = cacheUsage.CacheCreationInputTokens
resp.Usage.CacheReadInputTokens = cacheUsage.CacheReadInputTokens
if cacheProfile != nil {