Zero of fourteen models named Tailscale first on the direct prompt; one named Check Point Harmony SASE. Tailscale was named by twelve of the fourteen models and Check Point Harmony SASE by twelve and Tailscale carries 25 labels and Check Point Harmony SASE 30, so the shares are not directly comparable.
Named in three categories this edition.
Named in five categories this edition.
Share is the count of first choices across the direct, paraphrase, budget and scale prompts over all fourteen models, for a mid-market B2B company; rank is within the category; every quote names the model and the prompt it came from. Both figures come from the zero trust network access page.
Across every category in the October 2026 Edition, Tailscale and Check Point Harmony SASE were named in the same answer thirty-three times, of the 98 answers naming Tailscale and the 142 naming Check Point Harmony SASE. In those answers Check Point Harmony SASE took the first choice one time and Tailscale three.
| Model | Direct | Paraphrase | Comparative | Budget-constrained | Scale-constrained | Negative |
|---|---|---|---|---|---|---|
| Claude Haiku 4.5 | ||||||
| GPT-5.4 mini | ||||||
| Gemini 3.5 Flash | ||||||
| Perplexity Sonar | ||||||
| Grok 4.1 Fast | ||||||
| Mistral Small | ||||||
| DeepSeek V4 Flash | ||||||
| Llama 4 Maverick | ||||||
| Qwen 3.7 Flash | ||||||
| Kimi K2 | ||||||
| GLM 4.7 FlashX | ||||||
| MiniMax M2.5 | ||||||
| GPT-6 Luna | ||||||
| Muse Glimmer 30B |
Bold names in an answer are the products the judge labeled a first choice; a model naming several gives each of them that label. The full answer text for every row is in the record.
Every negative label with a quote, up to three, then the highest-weighted positives, up to three. Six of eight in this category shown.
“Good for developer-heavy environments, but operates more at network layer (Layer 4) than true application-layer zero trust” Kimi K2 · paraphrase prompt · soft negative
“may lack full enterprise ZTNA compliance (e.g., advanced posture or auditing) without self-hosting effort” Grok 4.1 Fast · budget prompt · soft negative
“License concerns: 2024 BSL (Business Source License) switch raised community concerns” MiniMax M2.5 · negative prompt · soft negative
“Pure ZTNA (e.g., Twingate, Tailscale, Cloudflare Zero Trust): Great if you just want to kill the VPN and grant app-level access. Low overhead, easy to implement.” Qwen 3.7 Flash · scale prompt · first choice
“Tailscale: For rapid deployment and ease of use (avoid if you need strict compliance granularities)” Qwen 3.7 Flash · negative prompt · first choice
“Best for developer-friendly private networking: Tailscale, for engineering teams that need fast, lightweight secure remote access.” Muse Glimmer 30B · paraphrase prompt · alternative
Every negative label with a quote, up to three, then the highest-weighted positives, up to three. Five of five in this category shown.
“Avoid or delay: Any vendor from the DEF CON research until you verify their remediation is architectural, not just patched” Kimi K2 · negative prompt · hard negative
“Hard-coded encryption keys (CVE-2025-3831) that exposed customer data through diagnostic logs” DeepSeek V4 Flash · negative prompt · hard negative
“When to avoid: If log security or crypto hygiene is critical.” Grok 4.1 Fast · negative prompt · hard negative
“Check Point Harmony SASE (formerly Perimeter 81) and Netskope ZTNA are the top recommendations” Mistral Small · direct prompt · first choice
“"Check Point Harmony SASE (formerly Perimeter 81) is a cloud-based zero-trust platform combining ZTNA with a Secure Web Gateway."” Muse Glimmer 30B · comparative prompt · alternative
Comparisons are drawn for the top eight products in each category, each against each. The output is the models' output; nothing here is a recommendation by the index.