Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion packages/backend/src/services/inspectors/usage-logging.ts
Original file line number Diff line number Diff line change
Expand Up @@ -10,9 +10,10 @@ import {
normalizeGeminiUsage,
normalizeOpenAIChatUsage,
normalizeOpenAIResponsesUsage,
extractUsageCostDetails,
} from '../../utils/usage-normalizer';
import { estimateKwhUsed } from '../inference-energy';
import { applyProviderReportedCost } from '../../utils/provider-cost';
import { applyProviderReportedCost, applyUsageCostDetails } from '../../utils/provider-cost';
import { DEFAULT_MODEL, DEFAULT_GPU_PARAMS } from '@plexus/shared';
import { recordQuotaUsage } from '../quota/quota-middleware';

Expand Down Expand Up @@ -149,6 +150,15 @@ export class UsageInspector extends PassThrough {
applyProviderReportedCost(this.usageRecord, reconstructed.providerReportedCost);
}

// Override with provider-reported cost from usage.cost_details if available
// Some providers include detailed cost breakdowns in the usage block
if (!this.usageRecord.providerReportedCost && reconstructed?.usage) {
const usageCostDetails = extractUsageCostDetails(reconstructed.usage);
if (usageCostDetails) {
applyUsageCostDetails(this.usageRecord, usageCostDetails);
}
}

// Use provider-reported energy if available, otherwise estimate
// Some providers emit `: energy {"energy_kwh": ...}` as SSE comments
if (reconstructed?.providerReportedEnergy?.energy_kwh != null) {
Expand Down
11 changes: 10 additions & 1 deletion packages/backend/src/services/response-handler.ts
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,8 @@ import { DebugLoggingInspector, UsageInspector } from './inspectors';
import { Readable } from 'stream';
import { DebugManager } from './debug-manager';
import { estimateKwhUsed } from './inference-energy';
import { applyProviderReportedCost } from '../utils/provider-cost';
import { applyProviderReportedCost, applyUsageCostDetails } from '../utils/provider-cost';
import { extractUsageCostDetails } from '../utils/usage-normalizer';
import { StallInspector, type StallConfig } from './inspectors/stall-inspector';
import { DEFAULT_GPU_PARAMS, DEFAULT_MODEL } from '@plexus/shared';
import type { GpuParams } from '@plexus/shared';
Expand Down Expand Up @@ -502,6 +503,14 @@ async function finalizeUsage(
if (reconstructed?.providerReportedCost) {
applyProviderReportedCost(usageRecord, reconstructed.providerReportedCost);
}

// Also check for cost_details in the usage block (some providers embed costs there)
if (!usageRecord.providerReportedCost && reconstructed?.usage) {
const usageCostDetails = extractUsageCostDetails(reconstructed.usage);
if (usageCostDetails) {
applyUsageCostDetails(usageRecord, usageCostDetails);
}
}
usageRecord.responseStatus = 'success';
usageRecord.durationMs = Date.now() - startTime;

Expand Down
Loading