| Category | Function | Share | Rank | Negative rate | Labels | Quadrant | Since September 2026 |
|---|---|---|---|---|---|---|---|
| SASE platforms | Network and edge | 2% | 9 of 45 | 43% | 53 | criticized challenger | ▲+2Since September 2026: 0% → 2%, +2 points. Inside the 11-point floor: within noise. Read over the models both editions asked. |
| Zero trust network access | Network and edge | 0% | 63 of 66 | 100% | 1 | under 10 labels · led by Twingate at 39% | =heldSince September 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise. Read over the models both editions asked. |
| Category | September 2026 | Now | Change | Reading | Rank |
|---|---|---|---|---|---|
| SASE | 0% | 2% | ▲+2Since September 2026: 0% → 2%, +2 points. Inside the 11-point floor: within noise. Read over the models both editions asked. | Within noise | Rank 11 → 9 of 45 |
| ZTNA | 0% | 0% | =heldSince September 2026: 0% → 0%, ±0 points. Inside the 11-point floor: within noise. Read over the models both editions asked. | Within noise | Rank 37 → 63 of 66 |
Shares here are read over the models both editions asked, so they can differ by a point or two from the standing above, which counts every model in this edition.
The floor is 11 points of share, measured: how far the models move a leader on their own when the same questions are asked twice with nothing changed. A larger change is movement; a smaller one is noise, and both are shown. Movement is read over the twelve models both editions asked; GPT-6 Luna, Muse Glimmer 30B joined this edition and are in the standing but not yet in the comparison. How the floor is measured · The editions
| Model | First choice | Alternative | Mention | Negative | Labels |
|---|---|---|---|---|---|
| Claude Haiku 4.5 | 0 | 0 | 1 | 1 | 2 |
| GPT-5.4 mini | 0 | 2 | 1 | 1 | 4 |
| Gemini 3.5 Flash | 0 | 1 | 0 | 5 | 6 |
| Perplexity Sonar | 0 | 0 | 1 | 2 | 3 |
| Grok 4.20 | 0 | 3 | 0 | 3 | 6 |
| Mistral Small | 1 | 1 | 1 | 1 | 4 |
| DeepSeek V4 Flash | 0 | 1 | 0 | 4 | 5 |
| Llama 4 Maverick | 0 | 2 | 3 | 0 | 5 |
| Qwen 3.7 Flash | 0 | 2 | 1 | 1 | 4 |
| Kimi K2 | 0 | 0 | 1 | 2 | 3 |
| GLM 4.7 FlashX | 0 | 0 | 1 | 1 | 2 |
| MiniMax M2.5 | 1 | 1 | 1 | 0 | 3 |
| GPT-6 Luna | 0 | 2 | 0 | 1 | 3 |
| Muse Glimmer 30B | 1 | 0 | 1 | 2 | 4 |
Verbatim evidence the judge attached to positive labels.
“Pros: Advanced threat protection, high scalability, strong integration with cloud environments” Mistral Small · SASE · comparative prompt · first choice
“Zscaler is the vendor most enterprises benchmark everyone else against.” Muse Glimmer 30B · SASE · comparative prompt · first choice
“Cloudflare One or Zscaler offer the best value” MiniMax M2.5 · SASE · budget prompt · first choice
“Choose Zscaler if you are moving 100% to the cloud and strictly enforcing Zero Trust principles.” Qwen 3.7 Flash · SASE · comparative prompt · alternative
Verbatim evidence attached to negative labels. A warning on a product with few labels is a warning; on a product with many, it is one voice among them.
“Frequent Outages: Over 2,855 tracked incidents over six years on Downdetector” GLM 4.7 FlashX · SASE · negative prompt · hard negative
“vendors like Zscaler or Cato may be worth considering—but they generally skew more enterprise and are less obviously budget-first” GPT-5.4 mini · SASE · budget prompt · soft negative
“they are often too complex, too expensive, and require too much dedicated staff to manage and fine-tune for a mid-market organization” Gemini 3.5 Flash · SASE · paraphrase prompt · soft negative
“Zscaler faced criticism regarding "bottlenecks" during peak usage times, attributed to a limited number of Points of Presence (PoPs)” Qwen 3.7 Flash · SASE · negative prompt · soft negative
Citations exist only for the models that return a source list, five of the fourteen in this edition, so these counts come from 154 of the 166 answers that named Zscaler Zero Trust Exchange and are not a share of its labels.
453 of the 453 domain citations in answers naming Zscaler Zero Trust Exchange came from somebody else's page.
Pages are listed as the models cited them.
Search figures are US estimates from DataForSEO, read September 28, 2026; AI search demand is its modeled, directional estimate, not a count of queries to any assistant. The answers are this edition's. Two measurements side by side: neither is read as the cause of the other.
An email the morning each edition publishes: where this product moved, where it held, and by how much against the noise floor. One address, confirmed by a click; a stop link in every email.
Already following? Everything you follow, with a stop for each.
What Zscaler Zero Trust Exchange's own pages state, read September 28, 2026: help.zscaler.com/zia/configuring-microsoft-exchange-zscaler-outbound-email-dlp. A claimed page can correct any of them.
Claiming is free and changes nothing in the data. A claimed page shows a verified contact who is told when each edition publishes and when Zscaler Zero Trust Exchange's standing changes by more than the noise floor; the right to propose corrections to the vendor table, meaning names the judge wrote that should or should not read as Zscaler Zero Trust Exchange, applied by version and listed in the change log; and a one-line description supplied by the vendor and marked as such.
A new claim receives the current edition's vendor brief for Zscaler Zero Trust Exchange by email, built from the raw record of the edition. It shows: