Model comparison
GLM-5 vs GLM-5.3
Compare published specifications, vendor API list prices and measured Borg performance. Both models are included in Borg.
GLM-5
GLM-5.3
Included in Borg
At a glance
| Specification | GLM-5 | GLM-5.3 |
|---|---|---|
| Provider | Z.ai | Z.ai |
| Release | 2026-02-12 Vendor source | 2026-08-18 Vendor source |
| Context | 200K tokens (vendor-stated) Vendor source | 1,048,576 tokens Vendor source |
| Modalities | Input: text; output: text | Input: text; output: text |
| Reasoning / thinking modes | Yes / Not published | Yes / Not published |
| Open weights | Yes | Yes |
| Vendor API list price / 1M tokens | $1 input / $3.2 output | $1.4 input / $4.4 output |
| Borg | Included in Borg subscription | Included in Borg subscription |
Z.aiMeasured on Borg
Metrics are not directly comparable to vendor benchmark claims. Models served different requests; this is observed gateway traffic, not a controlled head-to-head test.
Time to first token (gateway proxy)
No published measurement for this model.
Output tokens per second (whole turn)
| Metric | GLM-5 | GLM-5.3 |
|---|---|---|
| Time to first token (gateway proxy) p50 | Not published | 7293 ms (27495 samples) |
| Tokens per second (end to end) p50 | 33.3 output tokens per second (465 samples) | 25.2 output tokens per second (626315 samples) |
| Request success rate | Not published | Not published |
| Coding bench pass rate | Not published | 53% (15 samples) |
Measurement window, samples and methodology
The fleet reference is the median of published model medians, not a request-weighted result. Prompt lengths, reasoning settings and traffic differ between models.
GLM-5: 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.
GLM-5.3: Measured on Borg: SearchAtlas AI Gateway production request logs. 30-day window ending 2026-10-08; 659807 requests. Minimum 100 samples per published request metric.
Completion speed covers the whole turn including queueing and reasoning time; it is not a streaming tokens-per-second figure.
- 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
How to choose
GLM-5 is listed for multi-stage systems engineering, whereas GLM-5.3 is listed for complex coding-agent tasks. Choosing between GLM-5 and GLM-5.3 requires checking published specifications against observed Borg traffic for your workflow.
Key differences
| Spec | GLM-5 | GLM-5.3 |
|---|---|---|
| Context window | 200K tokens (vendor-stated) Vendor source | 1,048,576 tokens Vendor source |
| Reasoning modes | Yes / Not published | Yes / Not published |
| Release date | 2026-02-12 Vendor source | 2026-08-18 Vendor source |
| API price input / 1M | 1 Lower listed rate | 1.4 |
| API price output / 1M | 3.2 Lower listed rate | 4.4 |
| Time to first token (gateway proxy) | Not published | 7293 ms (27495 samples) |
| Output tokens per second (whole turn) | 33.3 output tokens/s (whole turn) (465 samples) Higher measured rate | 25.2 output tokens/s (whole turn) (626315 samples) |
| Request success | Not published | Not published |
| Coding bench | Not published | 53% (15 samples) |
FAQ
How do GLM-5 and GLM-5.3 compare for coding?
GLM-5 has a Borg coding benchmark of not published; GLM-5.3 has 53.33% pass rate across 15 samples and 3 tasks. The GLM-5–GLM-5.3 coding comparison cannot establish an overall winner from this small test alone.
How quickly do GLM-5 and GLM-5.3 start responding on Borg?
Borg median first response for GLM-5 is not published; for GLM-5.3 it is 7,293 ms. The GLM-5–GLM-5.3 gateway medians are observations, not guaranteed latency for individual requests.
What are the API prices for GLM-5 and GLM-5.3?
The listed API rate for GLM-5 is $1 input / $3.2 output per 1M tokens; for GLM-5.3 it is $1.4 input / $4.4 output per 1M tokens. Neither GLM-5 nor GLM-5.3 has its Borg subscription charge represented by those provider API rates. GLM-5 vendor source · GLM-5.3 vendor source
Explore both models
GLM-5 · GLM-5.3 · Borg pricing
Claude
Gemini