Activity & Logs

Trace GLM 5 API spend and failures from the account overview down to one request, with its model, key, tokens, credits, latency, and error code.

Open API Activity to inspect public API usage at request level.

Overview

Overview summarizes credits, requests, tokens, error rate, P95 latency, model usage, key usage, and recent error codes. Model, key, and error rows link directly into matching request logs.

Trends compares recent activity and lists daily requests, credits, tokens, errors, and P95 latency.

Explore

Explore uses the same bounded analytics query model as the Analytics API. Choose a metric (requests, credits, input_tokens, output_tokens, errors, or latency) and a dimension (day, model, key, or status).

Request logs

Filter Logs by Request ID, model, key, status, error code, or date range. Each request detail exposes:

  • Request ID
  • API key name/prefix
  • public model and upstream routing metadata
  • HTTP status and streaming mode
  • input/output tokens
  • estimated and actual credits
  • latency and time to first token when measurable
  • finish reason and error code

GLM5 does not store prompt or completion bodies in these request logs by default.

Request IDs

Public Chat Completions responses include X-Request-ID. Rate-limited and other authenticated failures also carry the ID, so support can correlate the client-visible failure with a request log when an account/key is known.