Z.ai · in Borg

GLM-5

GLM-5 begins Z.ai's next foundation-model generation after GLM-4.7, with a focus on agentic engineering rather than just coding completion.

Reasoning200K tokens (vendor-stated)text

200K tokens (vendor-stated)Context window

—Median time to first token (gateway proxy)

33.3Output tokens per second (whole turn)

$1 / $3.2Vendor API price in / out · per 1M

Included in Borg

Download Borg Borg for Business

At a glance

Specification	Value
Provider	Z.ai
Family	GLM
Release date	2026-02-12 Vendor source
Context window	200K tokens (vendor-stated) Vendor source
Input modalities	text
Output modalities	text
Reasoning	Yes
Open weights	Yes
Borg lane ids	`glm-5`, `z-ai/glm-5`
Versions in Borg	GLM-4.5 · GLM-4.6 · GLM-4.7 · GLM-5 · GLM-5.1 · GLM-5.2 · GLM-5.3
Median tokens per second (end to end) on Borg	33.3 output tokens per second

Not published by the vendor: thinking modes, knowledge cutoff, fusion roles.

What GLM-5 is good for

GLM-5 begins Z.ai's next foundation-model generation after GLM-4.7, with a focus on agentic engineering rather than just coding completion. It targets multi-stage software work: turning an underspecified goal into a plan, using tools, checking results and keeping architectural constraints in view. Z.ai also describes uses in scripts, roleplay and other extended writing, but the API model documented here accepts and returns text. It precedes GLM-5.1's explicit long-horizon optimization and GLM-5.2's much larger project-scale context. Use it for debugging difficult systems, developing full-stack features or testing autonomous agents where deeper planning is helpful. Borg includes GLM-5 with the subscription; Borg Fusion can choose it for a fitting step, and its direct lane is available when you want a fixed model.

multi-stage systems engineering
backend architecture and bug fixing
planning and checking agent actions
full-stack feature delivery

How Borg uses it

GLM-5 is available as a subscription catalog lane in Borg. The catalog does not publish which Fusion profiles route to this model, so no routing role is claimed here. You can pick it directly by setting the OpenAI-compatible gateway model field to a listed lane id. Check the current catalog for availability before building a workflow around a lane.

model: "glm-5"

Measured on Borg

Metric	Median	95th percentile	Samples
Time to first token (gateway proxy)	—	—	5
Tokens per second (end to end)	33.3 output tokens per second	74.6 output tokens per second	465

Output tokens per second (whole turn)

Higher is fasterFleet median 31.9 tokens/s

GLM-533.3 tokens/s

Time to first token (gateway proxy): time until the first streamed response event reaches the gateway; our closest available measure of time to first token
Tokens per second (end to end): tokens per second, end to end: provider-reported output tokens divided by the whole turn's elapsed time, including queueing and reasoning; not a streaming per-token rate

Completion speed covers the whole turn including queueing and reasoning time; it is not a streaming tokens-per-second figure.

Measured on Borg: SearchAtlas AI Gateway production request logs. 30-day window ending 2026-10-08; 566 requests. Minimum 100 samples per published request metric.

Benchmarks

Benchmark	Vendor-reported score	Source
SWE-bench Verified	77.8	Vendor source
Terminal Bench 2.0	56.2	Vendor source
GPQA-Diamond	86	Vendor source
AIME 2026 I	92.7	Vendor source

Pricing

Included in Borg subscription. Vendor API list price: $1 input / $3.2 output per 1M tokens. Vendor price source.

Compare subscription and API pricing · Cost Savings Calculator

FAQ

What is GLM-5?

GLM-5 is listed in Borg's subscription catalog as glm-5. GLM-5 begins Z.ai's next foundation-model generation after GLM-4.7, with a focus on agentic engineering rather than just coding completion.

Is GLM-5 in Borg?

Yes. GLM-5 is listed in Borg's subscription catalog under the lane glm-5. You can choose the lane directly, or explore the available Fusion profiles. Availability and model details can change; check the current catalog before deployment.

What context window does GLM-5 have?

The published context window is 200K tokens (vendor-stated). Consult the cited sources for details and current limits. Vendor source

Is GLM-5 good for coding?

Vendor-published evaluations appear in the benchmarks section; their settings may differ from your tasks. Its listed uses include multi-stage systems engineering, backend architecture and bug fixing, planning and checking agent actions, full-stack feature delivery. Try the lane for your own tasks, and compare the output with your requirements before choosing a default.

How fast is GLM-5 on Borg?