# Abnormal AI vs IRONSCALES: which do AI models recommend for email security, October 2026

IT AI Recommendation Index, October 2026 Edition, Email security. Two of fourteen models named Abnormal AI first on the direct prompt; one named IRONSCALES. Page: https://it-ai-index.com/security/email-security/abnormal-ai-vs-ironscales/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| Abnormal AI | 4% | #7 of 9 | 2% | 42 | 14 of 14 |
| IRONSCALES | 2% | #8 of 9 | 0% | 26 | 12 of 14 |

## The direct prompt, model by model

- DeepSeek V4 Flash: abnormal ai first (first choices: Abnormal AI) (alternatives: Check Point Harmony Email & Collaboration, Microsoft Defender for Office 365, Proofpoint Email Protection)
- GPT-6 Luna: abnormal ai first (first choices: Abnormal AI) (alternatives: Microsoft Defender for Office 365, Mimecast Email Security, Proofpoint Email Protection)
- Kimi K2: ironscales first (first choices: IRONSCALES) (alternatives: Abnormal AI, Mimecast Email Security, Proofpoint Email Protection)
- Claude Haiku 4.5: neither first, one named (first choices: Barracuda Email Protection) (alternatives: IRONSCALES, Mimecast Email Security, Proofpoint Email Protection)
- GPT-5.4 mini: neither first, one named (first choices: Microsoft Defender for Office 365) (alternatives: Abnormal AI, Mimecast Email Security, Proofpoint Email Protection)
- Gemini 3.5 Flash: neither first, one named (first choices: Check Point Harmony Email & Collaboration) (alternatives: Abnormal AI, Barracuda Email Protection, IRONSCALES)
- Grok 4.1 Fast: neither first, one named (first choices: Proofpoint 365 Total Protection) (alternatives: Abnormal AI, Mimecast Email Security)
- Qwen 3.7 Flash: neither first, one named (first choices: Proofpoint Email Protection) (alternatives: Abnormal AI, IRONSCALES, Microsoft Defender for Office 365)
- GLM 4.7 FlashX: neither first, one named (first choices: Microsoft Defender for Office 365) (alternatives: Abnormal AI, Barracuda Integrated Email Protection, IRONSCALES, Proofpoint Email Protection)
- Muse Glimmer 30B: neither first, one named (first choices: Proofpoint Email Protection) (alternatives: Abnormal AI, Barracuda Email Protection, Check Point Harmony Email & Collaboration, Guardian Digital, Microsoft Defender for Office 365, ThreatDown Email Security, Trustifi)
- Perplexity Sonar: neither named (first choices: Microsoft Defender for Office 365) (alternatives: Check Point Harmony Email & Collaboration, Mimecast Email Security)
- Mistral Small: neither named (first choices: Check Point Harmony Email & Collaboration, Proofpoint Email Protection) (alternatives: Mimecast Email Security, Symantec Email Security.cloud)
- Llama 4 Maverick: neither named
- MiniMax M2.5: neither named (first choices: Check Point Harmony Email & Collaboration, Mimecast Email Security)

## What the models said about Abnormal AI

- "Abnormal Security (caution, not avoid)" (DeepSeek V4 Flash, negative prompt, soft negative)
- "I'd shortlist Abnormal AI as the best add-on for catching business-email-compromise (BEC), vendor impersonation, and other attacks that slip past standard filtering" (GPT-6 Luna, direct prompt, first choice)
- "Abnormal Security is the top AI-native option for teams that need API-based detection of social-engineering attacks without gateway overhead." (Muse Glimmer 30B, comparative prompt, first choice)
- "These solutions (such as Abnormal Security, Avanan, or Sublime) connect directly to your email provider via native APIs." (Gemini 3.5 Flash, negative prompt, first choice)

## What the models said about IRONSCALES

- "prioritize Behavioral AI (like Abnormal AI or IRONSCALES) over signature-based gateways" (GLM 4.7 FlashX, negative prompt, first choice)
- "For most mid-market B2B companies, I'd recommend IRONSCALES as the top choice" (Kimi K2, direct prompt, first choice)
- "Modern solutions (like Abnormal Security, Check Point/Avanan, Ironscales, or Sublime Security) connect directly to Microsoft or Google via API." (Gemini 3.5 Flash, scale prompt, alternative)

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. Comparisons are drawn for the top eight products in each category. Published under CC BY 4.0; the output is the models' output, and nothing here is a recommendation by the index.