# WordPress vs Sanity: which do AI models recommend for Marketing CMS, September 2026

GTM AI Recommendation Index, September 2026 Edition, Content management systems for marketing sites. One of twelve models named WordPress first on the direct prompt; one named Sanity. Page: https://gtm-ai-index.com/marketing/marketing-cms/wordpress-vs-sanity/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| WordPress | 33% | #1 of 13 | 24% | 68 | 12 of 12 |
| Sanity | 6% | #5 of 13 | 7% | 27 | 11 of 12 |

## The direct prompt, model by model

- Claude Haiku 4.5: wordpress first (first choices: HubSpot Content Hub, WordPress) (alternatives: Contentful, Webflow)
- MiniMax M2.5: sanity first (first choices: Contentful, Sanity) (alternatives: Strapi, Webflow, WordPress)
- Gemini 3.5 Flash: neither first, one named (first choices: Webflow) (alternatives: HubSpot Content Hub, WordPress)
- Perplexity Sonar: neither first, one named (first choices: HubSpot Content Hub) (alternatives: Contentful, Sanity, Webflow, WordPress)
- Grok 4.1 Fast: neither first, one named (first choices: HubSpot Content Hub) (alternatives: Webflow, WordPress)
- Mistral Small: neither first, one named (first choices: Webflow) (alternatives: HubSpot Content Hub, WordPress)
- DeepSeek V4 Flash: neither first, one named (first choices: Webflow) (alternatives: HubSpot Content Hub, WordPress)
- Qwen 3.7 Flash: neither first, one named (first choices: HubSpot Content Hub, Webflow) (alternatives: Contentful, Sanity)
- Kimi K2: neither first, one named (first choices: Webflow) (alternatives: HubSpot Content Hub, Sanity, WordPress)
- GLM 4.7 FlashX: neither first, one named (first choices: HubSpot Content Hub, Webflow) (alternatives: Kentico Xperience, WordPress)
- GPT-5.4 mini: neither named (first choices: Contentful, Storyblok) (alternatives: Drupal)
- Llama 4 Maverick: neither named (first choices: HubSpot Content Hub)

## What the models said about WordPress

- "1. WordPress (Most Risky)
Why to avoid:" (Mistral Small, negative prompt, hard negative)
- "Managed Monolith (e.g., WordPress VIP, WP Engine) ... Trade-off: High maintenance overhead, security vulnerability surface area, and risk of "plugin bloat"" (Gemini 3.5 Flash, scale prompt, soft negative)
- "sources are blunt that it's a maintenance burden—plugin conflicts, security patches, and updates consume marketing-team time" (DeepSeek V4 Flash, paraphrase prompt, soft negative)
- "WordPress remains the undisputed king of budget marketing websites. Because the core software is free and open-source, your only fixed cost is web hosting." (Gemini 3.5 Flash, budget prompt, first choice)
- "WordPress is a popular open-source CMS that offers maximum flexibility and customization options, but requires more technical expertise" (Llama 4 Maverick, budget prompt, first choice)
- "WordPress for control... Best for: Flexible, SEO-heavy content sites, blogs, and custom marketing needs with massive plugin ecosystem" (Grok 4.1 Fast, comparative prompt, first choice)

## What the models said about Sanity

- "may be less suitable for certain organizations or use cases, such as ... Sanity, WordPress VIP" (Llama 4 Maverick, negative prompt, soft negative)
- "it strips away marketer autonomy... headless may slow you down." (Gemini 3.5 Flash, direct prompt, soft negative)
- "I'd suggest starting with Contentful or Sanity as they balance ease of use with enterprise features" (MiniMax M2.5, direct prompt, first choice)
- "Contentful/Sanity (Headless) | Dev-led marketing sites | Flexible APIs, fast scaling, marketer-friendly" (Grok 4.1 Fast, scale prompt, first choice)
- "Contentful / Sanity: Best "Headless" options. Very high cost of entry, but extremely scalable." (Qwen 3.7 Flash, scale prompt, first choice)

Share is the count of first choices across the direct, paraphrase, budget and scale prompts over all twelve 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.