# Radiant Security vs Torq: which do AI models recommend for AI SOC agents, October 2026

IT AI Recommendation Index, October 2026 Edition, AI SOC agents. Nine of fourteen models named Radiant Security first on the direct prompt; zero named Torq. Page: https://it-ai-index.com/security/ai-soc-agents/radiant-security-vs-torq/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| Radiant Security | 24% | #2 of 8 | 3% | 29 | 12 of 14 |
| Torq | 2% | #5 of 8 | 5% | 20 | 13 of 14 |

## The direct prompt, model by model

- Claude Haiku 4.5: radiant security first (first choices: Radiant Security) (alternatives: Conifers CognitiveSOC, Dropzone AI, Stellar Cyber Open XDR)
- Perplexity Sonar: radiant security first (first choices: Radiant Security) (alternatives: Dropzone AI, Stellar Cyber Open XDR)
- Grok 4.1 Fast: radiant security first (first choices: Dropzone AI, Prophet Security, Radiant Security) (alternatives: Arctic Wolf, Huntress Managed EDR, Microsoft Security Copilot)
- Mistral Small: radiant security first (first choices: Radiant Security, SentinelOne Purple AI)
- DeepSeek V4 Flash: radiant security first (first choices: Radiant Security, UnderDefense) (alternatives: Dropzone AI, Microsoft Security Copilot + Sentinel)
- Llama 4 Maverick: radiant security first (first choices: Radiant Security) (alternatives: SentinelOne Purple AI)
- Kimi K2: radiant security first (first choices: Radiant Security) (alternatives: Dropzone AI, Stellar Cyber Open XDR, UnderDefense Agentic AI SOC)
- GLM 4.7 FlashX: radiant security first (first choices: Radiant Security, UnderDefense MAXI AI SOC) (alternatives: Dropzone AI, Microsoft Security Copilot, Torq)
- Muse Glimmer 30B: radiant security first (first choices: Radiant Security, UnderDefense)
- Gemini 3.5 Flash: neither first, one named (first choices: Dropzone AI) (alternatives: Intezer AI SOC, Prophet Security, Radiant Security)
- Qwen 3.7 Flash: neither first, one named (first choices: SentinelOne Purple AI) (alternatives: Radiant Security)
- MiniMax M2.5: neither first, one named (first choices: UnderDefense MAXI AI SOC) (alternatives: Radiant Security)
- GPT-6 Luna: neither first, one named (first choices: Dropzone AI) (alternatives: Microsoft Security Copilot, Prophet Security, Torq)
- GPT-5.4 mini: neither named (first choices: Torq AI SOC Platform) (alternatives: CrowdStrike Falcon, Google Security Operations, Panther)

## What the models said about Radiant Security

- "Radiant is likely no longer a viable standalone choice for a typical mid-sized B2B stack." (Gemini 3.5 Flash, paraphrase prompt, hard negative)
- "Best for mid-market alert overload, effectively addressing the core pain point of too many alerts and not enough analysts" (Claude Haiku 4.5, direct prompt, first choice)
- "Radiant Security is praised for mid-market teams looking to reduce SIEM costs and alert noise." (Mistral Small, direct prompt, first choice)
- "Start with Dropzone AI, Prophet Security, or Radiant Security if building in-house capability" (Grok 4.1 Fast, direct prompt, first choice)

## What the models said about Torq

- "What to Avoid: "The Enterprise Giants" ... too heavy and expensive for a lean mid-market team." (Gemini 3.5 Flash, direct prompt, hard negative)

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.