As of November 18, 2026, no GLM-5.5 release is documented on the public Z.ai or BigModel pages reviewed for this article. The current documented family includes GLM-5.3, GLM-5.3-Flash and GLM-5.3-FlashX. This is a narrow statement about specific public sources, not a claim that Z.ai has never tested GLM-5.5 privately or will not release it later.
Review date: November 18, 2026. Scope: Z.ai's global model overview, release notes, pricing page and documentation index; the GLM-5.3 and GLM-5.3-Flash model pages; and the BigModel Chinese release log. We did not use authenticated API access, paid plans, private previews or non-public information. "Not listed" means not found in these pages on this date.
1. The short answer: no official GLM-5.5 release is documented
A release-date search should answer a precise question. The one answered here is: did the reviewed official pages publish a GLM-5.5 name, date, model ID or price? They did not.
| Official page | What it showed on November 18, 2026 | What it supports |
|---|---|---|
| Z.ai model overview | Latest Models: GLM-5.3 and GLM-5.3-Flash. No GLM-5.5 entry. | No current public model listing |
| Z.ai New Released | Latest entries: GLM-5.3-Flash on 2026-08-26 and GLM-5.3 on 2026-08-18. | No dated GLM-5.5 announcement |
| Z.ai pricing | Latest rows: GLM-5.3, GLM-5.3-Flash and GLM-5.3-FlashX. No GLM-5.5 row. | No published GLM-5.5 list price |
| Z.ai documentation index | Includes GLM-5.3 and GLM-5.3-Flash pages; no glm-5.5 page link. | No indexed GLM-5.5 model page |
| BigModel release log | Latest entry: GLM-5.3-Flash on 2026-08-26. No GLM-5.5 entry. | No dated GLM-5.5 announcement on the China platform log |
The distinction matters. A model can be tested internally, offered to selected partners or planned for a future date without appearing in a public catalog. Those possibilities are outside this review. The evidence supports "no public official release was found in the sources checked," not "the model does not exist" and not "the model will never ship."
2. The specific unknowns
For an integration team, the missing item is not just a date. A usable release needs a contract. As of the review date, none of the following was published for GLM-5.5 in the pages checked:
| Item | Status in reviewed sources | Consequence |
|---|---|---|
| Official release date | Not published | Third-party dates cannot be used as a launch schedule |
| Model ID and API endpoint | No documented glm-5.5 identifier | A guessed ID may fail or behave differently by provider |
| Context window and maximum output | Not documented for GLM-5.5 | GLM-5.3 limits cannot be inherited |
| Input and output modalities | Not documented for GLM-5.5 | Text-only versus multimodal changes application design |
| Pricing, cached-input rate and billing unit | No pricing row | Budgets cannot be projected from GLM-5.3 rates |
| Model card, license and weight availability | Not found in the reviewed pages | Open-weight assumptions need an official artifact |
| Rate limits, regions and account eligibility | Not documented for GLM-5.5 | Availability may differ by platform and region |
| ToAPIs availability and price | Not established by this article | Vendor documentation does not define a gateway catalog |
Community screenshots, anonymous benchmark fragments, repository names and third-party "release date" articles can be useful leads. They do not establish a public model contract. This article does not reproduce their scores or treat their specifications as confirmed. A placeholder page, a filename, or a CLI accepting a string is not a release announcement.
3. The documented baseline: GLM-5.3 and GLM-5.3-Flash
If you are evaluating a GLM-family model with a published contract, the current public baseline in the official documentation is GLM-5.3, GLM-5.3-Flash and GLM-5.3-FlashX. The table below records only what the official pages state. It is not a forecast for GLM-5.5.
| Item | GLM-5.3 | GLM-5.3-Flash |
|---|---|---|
| Official release note | 2026-08-18 on Z.ai; the BigModel log shows 2026-08-19 | 2026-08-26 on both release logs |
| Context / maximum output | 1M / 128K tokens | 1M / 128K tokens |
| Input / output | Text input, text output | Video, image, text and file input; text output |
| Reasoning controls | Always enabled; low, high, max; default max | Same text parameters; thinking cannot be disabled |
| Z.ai list price per 1M tokens | $1.40 input / $0.26 cached input / $4.40 output | $0.15 input / $0.03 cached input / $0.50 output |
Z.ai also lists GLM-5.3-FlashX at $0.37 input, $0.075 cached input and $1.25 output per 1M tokens, with a documented inference speed of about 200 tokens per second. It belongs to the current family, not to GLM-5.5.
These are Z.ai's global USD list prices. They are not a ToAPIs quote, a domestic CNY price, or a promise of availability in any particular account. The GLM-5.3 documentation also contains a migration constraint worth noting: thinking.type only supports enabled, and an application that currently sends disabled must change that value and set reasoning_effort to a supported level such as low before switching to glm-5.3. That is a GLM-5.3 fact. It says nothing about GLM-5.5's eventual parameters.
4. How to prepare without guessing a contract
The useful work before a release is to make the integration boundary explicit. Do not hardcode a guessed glm-5.5 identifier in production, and do not assume that a new version inherits every field from GLM-5.3.
- Keep the candidate disabled. Put any unverified model name outside the production allowlist until an official model page and request schema exist.
- Separate capability from vendor. Store model ID, endpoint, context limit, modality, reasoning options and pricing as configuration, not as assumptions embedded in application code.
- Freeze an evaluation pack. Include structured output, tool calling, long-context retrieval, refusal behavior and any multimodal task your product needs. Define acceptance rules before testing.
- Verify identity at the provider. Record the requested ID, the returned model field, the account scope, the request date and the provider's documentation. A successful call is evidence for that account and route, not a general availability guarantee.
- Measure the real billing unit. Check whether the provider charges per token, per request, per image or with a minimum. Include cached input, reasoning tokens, failed attempts and retries where applicable.
- Keep a rollback path. A new model should not be the only route for a user-visible feature. Define stop conditions for errors, cost overruns and quality regressions.
A launch checklist is only useful if it can fail. If the provider cannot supply a documented model ID, a supported request shape and a billing rule, the correct state is "not ready," even if a demo looks convincing.
To check what ToAPIs currently lists, use the model catalog, pricing page and API documentation. These are verification entry points, not a promise that GLM-5.5, GLM-5.3 or any other model is available on a particular account.
5. What would change this answer
This page should be updated only when one of these official signals appears:
- A dated Z.ai or BigModel release note that names GLM-5.5.
- An official model page or model card with a model ID, context window, output limit, modalities and license.
- A pricing row with the billing unit, region and effective date.
- API reference documentation covering endpoints, parameters, limits and error behavior.
- A reproducible authenticated request that returns a documented identifier, together with provider confirmation of the route. That would establish availability for a specific account and route, not general availability by itself.
- An official correction or clarification that changes the status described here.
This article is a dated snapshot, not a release calendar. It does not promise weekly checks or automatic updates. Before reusing the conclusion, reopen the official pages and record the date and URL.
6. Frequently asked questions
Has GLM-5.5 been released?
No public official release naming GLM-5.5 appeared in the Z.ai or BigModel pages reviewed on November 18, 2026. That finding is limited to those public sources. It does not rule out private testing, partner access or an unannounced future release.
What is the GLM-5.5 release date?
No official release date is published in the sources checked. Third-party dates, countdown pages and social posts are not official confirmations.
What are the GLM-5.5 context window, modalities and price?
They are not documented in the official pages reviewed. Any specific context size, token limit, benchmark score or price attributed to GLM-5.5 is unverified until an official model card, API reference or pricing row appears. Do not copy GLM-5.3 values into a GLM-5.5 budget as if they were confirmed.
What should I use while GLM-5.5 is unconfirmed?
The documented baseline includes GLM-5.3, GLM-5.3-Flash and GLM-5.3-FlashX. Verify the exact provider, region, model ID, limits and price before integrating. This article does not establish ToAPIs availability or routing for these models or for GLM-5.5.
Does a screenshot or third-party benchmark prove that GLM-5.5 exists?
No. It can be a lead for further verification, but it does not establish a public model identity, request contract, price or availability. Treat anonymous benchmarks and unverified screenshots as claims, not measurements.
What would update this article?
An official release note, model page, model card or license, pricing entry, API reference or official correction. The exact triggers are listed in section 5. No fixed update cadence is promised.
7. Sources and review scope
- Z.ai: Models overview — current model list; accessed 2026-11-18.
- Z.ai: New Released — release notes; accessed 2026-11-18.
- Z.ai: Pricing — USD list prices; accessed 2026-11-18.
- Z.ai: Documentation index — model page index; accessed 2026-11-18.
- Z.ai: GLM-5.3 — context, output, reasoning and migration details; accessed 2026-11-18.
- Z.ai: GLM-5.3-Flash/FlashX — modalities, limits and model codes; accessed 2026-11-18.
- BigModel: Release log — China platform release history; accessed 2026-11-18.
- BigModel: GLM-5.3 — Chinese model documentation; accessed 2026-11-18.
All sources were public at the time of review. Prices, dates and availability can change. This review does not establish a private preview, a future release, a ToAPIs route or a guaranteed production result.







