# Abnormal AI: how AI models rank it, November 2026

Gane Index, November 2026 Edition. Named in 45 judge labels across 2 categories by 14 of 14 models. Page: https://www.gane.ai/index/it/vendors/abnormal-ai

## Standing by category

| Category | Share | Rank | Negative rate | Labels |
|---|---|---|---|---|
| [Email security](https://www.gane.ai/index/it/security/email-security/index.md) | 8% | 2 | 9% | 43 |
| [Identity threat detection and response](https://www.gane.ai/index/it/identity/identity-threat-detection-and-response/index.md) | 0% | 48 | 0% | 2 |

## What the models said for it

- "Add Abnormal AI (or an equivalent AI-based layer) to catch the sophisticated BEC and social engineering threats that generic gateways miss" (DeepSeek V4 Flash, Email security)
- "Consider modern platforms like Abnormal Security, Material Security, or Microsoft Defender for Office 365" (DeepSeek V4 Flash, Email security)
- "My default pick: Abnormal AI for a mid-market B2B company using Microsoft 365 or Google Workspace" (GPT-6 Luna, Email security)
- "Consider Abnormal Security or Avanan. API integrations avoid MX record routing complexity" (Gemini 3.5 Flash, Email security)

## And against it

- "Avoid for Budget Scenarios - Abnormal Security - ... pricing starts at $25,000/year" (MiniMax M2.5, Email security)
- "Abnormal Security — powerful but requires expensive add-ons" (DeepSeek V4 Flash, Email security)
- "without the complexity or high costs of tools like Abnormal Security (which starts at $25K+/year)" (Grok 4.20, Email security)
- "Caveat for mid-market: Pricing starts around $35–$60/mailbox/year with minimum commitments" (Kimi K2, Email security)

## By model

| Model | First choice | Alternative | Mention | Warning | Labels |
|---|---|---|---|---|---|
| Claude Haiku 4.5 | 0 | 0 | 1 | 0 | 1 |
| GPT-5.4 mini | 0 | 1 | 0 | 0 | 1 |
| Gemini 3.5 Flash | 3 | 1 | 1 | 0 | 5 |
| Perplexity Sonar | 0 | 3 | 1 | 0 | 4 |
| Grok 4.20 | 0 | 2 | 1 | 1 | 4 |
| Mistral Small | 0 | 0 | 1 | 0 | 1 |
| DeepSeek V4 Flash | 2 | 3 | 0 | 1 | 6 |
| Llama 4 Maverick | 0 | 0 | 1 | 0 | 1 |
| Qwen 3.7 Flash | 0 | 2 | 0 | 0 | 2 |
| Kimi K2 | 0 | 3 | 2 | 1 | 6 |
| GLM 4.7 FlashX | 0 | 1 | 2 | 0 | 3 |
| MiniMax M2.5 | 0 | 2 | 2 | 1 | 5 |
| GPT-6 Luna | 1 | 2 | 0 | 0 | 3 |
| Muse Glimmer 30B | 0 | 3 | 0 | 0 | 3 |

## Sources returned with these answers

The sites and pages cited by the answers that named Abnormal AI, from 107 of the 117 answers that named it (only the models that return a source list).

### Domains cited

- checkpoint.com: 52
- g2.com: 41
- huntress.com: 40
- microsoft.com: 40
- securityboulevard.com: 34
- proofpoint.com: 32
- expertinsights.com: 29
- mimecast.com: 24
- costbench.com: 22
- abnormal.ai: 20 (its own site)

### Pages cited

- https://huntress.com/business-email-compromise-guide/10-best-email-security-providers: 31
- https://checkpoint.com/comparison/best-email-security-companies: 25
- https://proofpoint.com/us/compare/email-security-solutions: 21
- https://checkpoint.com/cyber-hub/threat-prevention/what-is-email-security/best-alternative-to-microsoft-defender: 20
- https://microsoft.com/en-us/security/business/siem-and-xdr/microsoft-defender-office-365: 19
- https://viewpointanalysis.com/post/email-security-software-options-2026: 17
- https://expertinsights.com/email-security/top-email-security-gateways: 16
- https://guideflow.com/blog/email-security-software: 15

## In its own words

What Abnormal AI's own pages state, read 2026-09-28. Stated by the vendor, not checked by the index.

- Positioning: provides a leading cloud-native email security platform that leverages AI-based behavioral data science to stop socially-engineered and never-seen-before email attacks that evade traditional secure email gateways (https://abnormal.ai/profile/abnormal-ai)
- For: finance teams (https://abnormal.ai/learning/invoice-fraud-ai)
- Customers named: Valvoline

## Record

- Method: https://www.gane.ai/index/it/methodology
- License: CC BY 4.0. Cite as Gane Index, November 2026 Edition, www.gane.ai/index/it.