| Category | Function | Share | Rank | Negative rate | Labels | Quadrant | Since September 2026 |
|---|---|---|---|---|---|---|---|
| Zero trust network access | Network and edge | 3% | 6 of 66 | 45% | 51 | criticized challenger | ▲+1Since September 2026: 2% → 4%, +1 point. Inside the 11-point floor: within noise. Read over the models both editions asked. |
| SASE platforms | Network and edge | 2% | 8 of 45 | 0% | 2 | under 10 labels · led by Cato Networks at 35% | ▲+2Since September 2026: 0% → 2%, +2 points. Inside the 11-point floor: within noise. Read over the models both editions asked. |
| Microsegmentation | Network and edge | 2% | 9 of 94 | 0% | 14 | accepted challenger |
| 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 34 → 8 of 45 |
| Microsegmentation | 0% | 2% | newNew since September 2026: not ranked then, 2% now. | New this edition | Rank 9 of 94, unchanged |
| ZTNA | 2% | 4% | ▲+1Since September 2026: 2% → 4%, +1 point. Inside the 11-point floor: within noise. Read over the models both editions asked. | Within noise | Rank 6 of 66, unchanged |
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 | 1 | 2 | 3 | 0 | 6 |
| GPT-5.4 mini | 2 | 1 | 1 | 0 | 4 |
| Gemini 3.5 Flash | 1 | 0 | 0 | 4 | 5 |
| Perplexity Sonar | 1 | 1 | 1 | 1 | 4 |
| Grok 4.20 | 1 | 3 | 0 | 3 | 7 |
| Mistral Small | 3 | 2 | 1 | 0 | 6 |
| DeepSeek V4 Flash | 1 | 0 | 0 | 4 | 5 |
| Llama 4 Maverick | 0 | 1 | 2 | 0 | 3 |
| Qwen 3.7 Flash | 0 | 3 | 0 | 2 | 5 |
| Kimi K2 | 1 | 1 | 0 | 3 | 5 |
| GLM 4.7 FlashX | 1 | 0 | 1 | 3 | 5 |
| MiniMax M2.5 | 0 | 3 | 1 | 1 | 5 |
| GPT-6 Luna | 0 | 1 | 2 | 0 | 3 |
| Muse Glimmer 30B | 0 | 1 | 1 | 2 | 4 |
Verbatim evidence the judge attached to positive labels.
“Zscaler Private Access is the market-leading ZTNA solution, delivering inside-out connectivity that never exposes applications to the internet” Claude Haiku 4.5 · ZTNA · comparative prompt · first choice
“Complex, global enterprise needing high compliance | Zscaler Private Access | Robust, highly certified, and built for massive scale.” Gemini 3.5 Flash · ZTNA · comparative prompt · first choice
“Zscaler is often treated as the canonical pure-play ZTNA vendor... Choose Zscaler if you want a widely recognized ZTNA leader” GPT-5.4 mini · ZTNA · comparative prompt · first choice
“The global leader in SASE (Secure Access Service Edge). ZPA is the dominant choice for large enterprises and government agencies” GLM 4.7 FlashX · ZTNA · comparative prompt · first choice
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.
“Avoid Zscaler for mid-market: It's powerful but designed for Fortune 500 scale—complex, expensive, and often overkill” Kimi K2 · ZTNA · paraphrase prompt · hard negative
“ZTNA Vendors to Approach with Caution ... Zscaler ... CVE-2025-54982: A SAML authentication bypass” DeepSeek V4 Flash · ZTNA · negative prompt · hard negative
“CVE-2025-54982: SAML authentication bypass due to failure to validate signed assertions” Kimi K2 · ZTNA · negative prompt · hard negative
“When to avoid: Small/mid-sized businesses or cost-sensitive setups” Grok 4.20 · ZTNA · negative prompt · hard negative
Citations exist only for the models that return a source list, five of the fourteen in this edition, so these counts come from 195 of the 209 answers that named Zscaler Private Access and are not a share of its labels.
488 of the 531 domain citations in answers naming Zscaler Private Access 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 Private Access's own pages state, read September 28, 2026: help.zscaler.com/zpa/about-private-clouds, help.zscaler.com/zpa/editing-private-clouds, help.zscaler.com/zpa/configuring-private-clouds, help.zscaler.com/zpa/about-private-access-roles, help.zscaler.com/zpa/what-zscaler-private-access. 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 Private Access'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 Private Access, 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 Private Access by email, built from the raw record of the edition. It shows: