This guide uses source checks from Sep 25, 2026. Provider and gateway prices can change; preserve the checked date with every forecast.
Why This Topic Matters Now
The Sep 24 market sweep kept `glm api` in the active query set, while the latest AIWave posts already covered GLM output budgets and tool-calling contract tests. The next useful angle is evaluation design: which task slices justify a GLM-5 rollout, how should agent paths be scored, and which failures must block promotion? A model that writes a good final answer can still be unsafe if it selects the wrong tool or ignores a stop condition.
This matrix is written for Tier 1 and Tier 2 teams comparing GLM-5 against an existing route in an OpenAI-compatible stack. It separates provider onboarding and pricing context from AIWave live route evidence and dated gateway rates. The recommended result is a small, versioned evaluation set with explicit quality, cost, tool, privacy, and rollback gates—not a permanent ranking.
Source Facts Checked Today
AIWave /api/pricing was checked from production on Sep 25, 2026 and returned HTTP 200, success=true, 73 live route rows, pricing_version a42d372ccf0b5dd13ecf71203521f9d2, auto_groups=['default'], group_ratio default=1 and vip=0.9. The public /api/v1/pricing endpoint returned HTTP 200 with 56 dated USD rows, pricing_version 83f77abde81ee3a096a672ed959ccc096f5d37a45c177ae8e03229456b5415a5, checked=2026-09-10, and updated_at=2026-09-18. Use the live response for route availability and the dated JSON for a forecast; they are not one interchangeable rate table.
Z.AI's quick-start page checked on Sep 25, 2026 lists GLM-5.3 and GLM-5.3-FLASH among the platform's model choices and documents HTTP, Python SDK, Java SDK, and OpenAI SDK compatibility paths. This supports an integration check; it does not prove that a GLM-5 route in another gateway exposes the same model version or feature set.
The live AIWave route response checked on Sep 25, 2026 includes glm-5 and glm-5-turbo with OpenAI endpoint metadata. The live response establishes current gateway route information only. Keep agent feature claims tied to the exact model ID and acceptance evidence rather than inferring them from the family name.
The dated AIWave public pricing JSON checked in this run lists glm-5 at $1.55 input, $0.40000075 cache-hit input, and $4.96 output per 1M tokens, effective 2026-08-27; glm-5-turbo is listed at $1.80 input, $0.4800006 cache-hit input, and $5.40 output. These are dated gateway base-rate rows and should not be merged with the dynamic Z.AI provider pricing page.
Planning Matrix
A source-dated planning matrix keeps the page useful for engineers and procurement reviewers. It turns a search query into an auditable route decision instead of a loose model preference.
| Evaluation slice | Blocker to watch | Promotion evidence |
|---|---|---|
| Reasoning | Plausible but unsupported conclusion | Rubric score and citation fixture |
| Tool selection | Wrong or unnecessary tool call | Allowlist and trace |
| Arguments | Schema-valid but unsafe values | Semantic validator and rejection |
| State control | Agent continues after stop | Turn ceiling and stop reason |
| Budget | Output or retries dominate spend | Token ledger and attempt count |
| Rollback | Route change breaks an app contract | Previous model ID and fixture hash |
Implementation Pattern
The implementation pattern keeps credentials as placeholders, pins the AIWave base URL, records the model, and leaves room for route-specific controls. Production applications should move credentials into environment or secret storage.
from openai import OpenAI
client = OpenAI(
api_key="YOUR_API_KEY_HERE",
base_url="https://aiwave.live/v1",
)
policy = {
"model": "glm-5",
"task_slice": "bounded-release-review",
"max_tool_turns": 2,
"max_tokens": 320,
"checked_at": "2026-09-25",
}
result = client.chat.completions.create(
model=policy["model"],
messages=[{"role": "user", "content": "Return one risk and one reversible next step."}],
temperature=0.0,
max_tokens=policy["max_tokens"],
)
print({"policy": policy, "finish": result.choices[0].finish_reason,
"usage": result.usage, "request_id": getattr(result, "id", None)})
Turn the Query Into a Contract
For a GLM-5 agent evaluation and rollback matrix, define the request shape, model ID, data class, output ceiling, timeout, retry ceiling, owner, and source date before the first trial. A short contract gives engineering, security, and finance the same object to review when a provider changes a route or billing field.
Separate Live Routes From Dated Rates
The live AIWave pricing response answers which route rows and endpoint types are available at check time. The public pricing JSON is a dated USD snapshot for forecasting. Store both URLs, versions, checked dates, model IDs, and account-group context instead of presenting a volatile source as a permanent quote.
Use a Small Acceptance Set
A useful canary covers a normal request, a malformed request, a repeated prefix, a long output, a disconnect, and a deliberate stop condition. Record request ID, model ID, status, token usage, finish reason, retry count, and reviewer outcome. This turns a search result into evidence that can survive a route update.
Keep Data and Credentials Bounded
OpenAI-compatible clients reduce integration work, but they do not choose the right data boundary. Keep the credential server-side, use a visible placeholder in examples, redact fixtures, and attach a data-class decision to every route policy. Do not let a feature flag or model alias silently widen what crosses the API.
Make Recovery Observable
Retry only failures that are safe to retry and cap every fallback. Preserve the original request ID, mark the stop reason, and distinguish provider errors from client validation, policy rejection, and budget stops. Silent loops hide both reliability failures and billing variance.
Use AIWave's Evidence Layer
Use the Models docs, Chat Completions docs, dated Pricing JSON, and Status. Recheck the live route table before rollout, the dated pricing JSON before a budget review, the status page before a launch window, and the trust page before procurement. Keep each checked date visible in the record.
Release Gate
Promotion is ready when the provider source is dated, the AIWave route is rechecked, the acceptance set passes, the billing fields are understood, and a named owner can stop or reverse the change. If a field is unknown, label the work as a trial rather than production.