fix(billing): correct tiered billing settlement and edge cases

- quota.go: add missing SettleBilling call in PostWssConsumeQuota
- text_quota.go: gate InjectTieredBillingInfo on tieredBillingApplied bool
  instead of tieredResult != nil, so fallback billing still logs metadata
- price.go: remove quotaBeforeGroup == 0 from freeModel condition to avoid
  bypassing settlement for output-only expressions
- tiered_settle.go: split cc/cc1h subtraction using UsageSemantic to
  distinguish OpenAI vs Claude cache creation token formats
- pricing.go: only set BillingMode when a non-empty expression exists
- useModelPricingEditorState.js: only write billing_mode when
  finalBillingExpr is non-empty

service/text_quota.go

@@ -330,6 +330,7 @@ func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
 	summary := calculateTextQuotaSummary(ctx, relayInfo, usage)
 
 	var tieredResult *billingexpr.TieredResult
+	tieredBillingApplied := false
 	if originUsage != nil {
 		var tieredUsedVars map[string]bool
 		if snap := relayInfo.TieredBillingSnapshot; snap != nil {
@@ -337,6 +338,7 @@ func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
 		}
 		tieredOk, tieredQuota, tieredRes := TryTieredSettle(relayInfo, BuildTieredTokenParams(usage, summary.IsClaudeUsageSemantic, tieredUsedVars))
 		if tieredOk {
+			tieredBillingApplied = true
 			tieredResult = tieredRes
 			summary.Quota = composeTieredTextQuota(relayInfo, summary, tieredQuota, tieredRes)
 		}
@@ -451,7 +453,7 @@ func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
 		// prompt/cache fields here, otherwise old upstream payloads may be double-counted.
 		other["input_tokens_total"] = usage.InputTokens
 	}
-	if tieredResult != nil {
+	if tieredBillingApplied {
 		InjectTieredBillingInfo(other, relayInfo, tieredResult)
 	}