Seven of fourteen models named Human Interest first on the direct prompt; two named ADP Retirement Services. Human Interest was named by thirteen of the fourteen models and ADP Retirement Services by fourteen and Human Interest carries 42 labels and ADP Retirement Services 41, so the shares are not directly comparable.
Named in one category this edition.
Named in one category 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 401(k) providers page.
Across every category in the October 2026 Edition, Human Interest and ADP Retirement Services were named in the same answer sixty-six times, of the 105 answers naming Human Interest and the 128 naming ADP Retirement Services. In those answers ADP Retirement Services took the first choice six times and Human Interest sixteen.
| Model | Direct | Paraphrase | Comparative | Budget-constrained | Scale-constrained | Negative |
|---|---|---|---|---|---|---|
| Claude Haiku 4.5 | ||||||
| GPT-5.4 mini | AR | |||||
| Gemini 3.5 Flash | ||||||
| Perplexity Sonar | AR | AR | ||||
| Grok 4.1 Fast | ||||||
| Mistral Small | AR | AR | AR | |||
| DeepSeek V4 Flash | ||||||
| Llama 4 Maverick | ||||||
| Qwen 3.7 Flash | AR | |||||
| Kimi K2 | AR | |||||
| GLM 4.7 FlashX | AR | AR | ||||
| MiniMax M2.5 | ||||||
| GPT-6 Luna | AR | |||||
| Muse Glimmer 30B | AR |
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 seven in this category shown.
“slam poor support, payroll delays/errors, and communication issues” Grok 4.1 Fast · negative prompt · hard negative
“Human Interest, Slavic401k, Guideline, and Ubiquity/Online 401(k) in various user reviews or forum posts reporting high costs, poor service” Perplexity Sonar · negative prompt · soft negative
“notable player, but it has historically been more focused on smaller businesses than true mid-market complexity” GPT-5.4 mini · direct prompt · soft negative
“Best Tech-Forward & Automated: Human Interest ... If you use Rippling, Gusto, or BambooHR and want the lowest administrative burden: Go with Human Interest.” Gemini 3.5 Flash · direct prompt · first choice
“Human Interest is recommended as the best fit for mid-market employers that want managed 401k administration with structured onboarding and conversion support.” Claude Haiku 4.5 · direct prompt · first choice
“Human Interest: Best Overall Value & Ease of Use ... go with Human Interest or Guideline” Gemini 3.5 Flash · budget prompt · first choice
Every negative label with a quote, up to three, then the highest-weighted positives, up to three. Six of eight in this category shown.
“ADP's fee structures can sometimes be less transparent than modern digital-first providers ... their underlying fund options may feature slightly higher expense ratios.” Gemini 3.5 Flash · paraphrase prompt · soft negative
“Large, Bundled Payroll Providers Common Examples: *ADP Retirement Services and Paychex*... historically carry high administrative and recordkeeping costs” Gemini 3.5 Flash · negative prompt · soft negative
“Concerns about "bundled" packages that may inflate total costs and obscure fee structures.” GLM 4.7 FlashX · negative prompt · soft negative
“This is widely considered the best option for very small teams because its pricing structure is designed to minimize costs” Qwen 3.7 Flash · budget prompt · first choice
“Shortlist: ADP Enhanced (mid-sized focus), Fidelity, Vanguard, Schwab, or Human Interest/ForUsAll for tech-forward.” Grok 4.1 Fast · scale prompt · first choice
“ADP and Fidelity are often the top recommendations due to their flexibility, strong service, and tailored solutions” Mistral Small · paraphrase prompt · first choice
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.