# Mercury vs Novo: which do AI models recommend for business banking, October 2026

Finance AI Recommendation Index, October 2026 Edition, Business banking. One of fourteen models named Mercury first on the direct prompt; zero named Novo. Page: https://finance-ai-index.com/treasury/business-banking/mercury-vs-novo/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| Mercury | 23% | #1 of 14 | 18% | 44 | 14 of 14 |
| Novo | 10% | #4 of 14 | 5% | 22 | 11 of 14 |

## The direct prompt, model by model

- GLM 4.7 FlashX: mercury first (first choices: Mercury) (alternatives: Airwallex, Bank of America, Brex, Chase Business Banking)
- Gemini 3.5 Flash: neither first, one named (first choices: Rho) (alternatives: Bank of America, Brex, Chase Business Banking, Mercury, Ramp)
- Qwen 3.7 Flash: neither first, one named (first choices: Ramp) (alternatives: Aspire, Bank of America, Chase Business Banking, Mercury, Oxygen Banking)
- Kimi K2: neither first, one named (first choices: Chase Performance Business Checking) (alternatives: Brex, Mercury, Relay)
- Claude Haiku 4.5: neither named (first choices: Chase Business Banking) (alternatives: Bluevine Premier Business Checking, KeyBank, Relay)
- GPT-5.4 mini: neither named (first choices: J.P. Morgan Commercial Banking) (alternatives: Bank of America, Wells Fargo Commercial Banking)
- Perplexity Sonar: neither named (first choices: Rho) (alternatives: Airwallex, Brex, Capital One Enhanced Checking)
- Grok 4.1 Fast: neither named (first choices: Chase Business Banking) (alternatives: Bank of America, Capital One)
- Mistral Small: neither named (first choices: Capital One) (alternatives: PNC Bank Treasury Enterprise Plan, Truist Bank, Wells Fargo)
- DeepSeek V4 Flash: neither named (first choices: Bank of America, Chase Business Banking) (alternatives: Capital One)
- Llama 4 Maverick: neither named
- MiniMax M2.5: neither named (first choices: PNC Bank) (alternatives: Capital One, First Internet Bank, Rho, Wise Business)
- GPT-6 Luna: neither named (first choices: J.P. Morgan Commercial Banking) (alternatives: Bank of America, Citibank, U.S. Bank)
- Muse Glimmer 30B: neither named (first choices: Bank of America, Chase Business Banking)

## What the models said about Mercury

- "Mercury, Relay, and BlueVine draw approximately four times the complaints of other business banking providers" (MiniMax M2.5, negative prompt, hard negative)
- "Highest rate of negative reviews among business banks (15.4/month), with ~41% citing sudden account closures" (Grok 4.1 Fast, negative prompt, hard negative)
- "Accounts are being closed or frozen at an unusually high rate ... Highest-Caution Fintech Providers" (DeepSeek V4 Flash, negative prompt, hard negative)
- "For a US startup with a tiny budget: Novo or Mercury (both completely free with $0 minimums and excellent features)." (DeepSeek V4 Flash, budget prompt, first choice)
- "Bluevine or Mercury currently offer the best combination of zero fees, features, and accessibility" (MiniMax M2.5, budget prompt, first choice)
- "Mercury has effectively become the default operating account for mid-market B2B companies." (GLM 4.7 FlashX, direct prompt, first choice)

## What the models said about Novo

- "Online platforms like Novo or Lili are great for lean startups, but they lack a physical branch presence." (Gemini 3.5 Flash, negative prompt, soft negative)
- "Zero-cost digital first: Novo or American Express Business Checking if you want 100% free with no minimums and no fees on ACH/wires." (Muse Glimmer 30B, budget prompt, first choice)
- "Opt for established options like Novo, Lili, or traditional banks (e.g., Wells Fargo) with fewer closure complaints." (Grok 4.1 Fast, negative prompt, first choice)
- "If you want a simple, fee‑free online account: Go with Novo, Bluevine, or Lili." (GLM 4.7 FlashX, budget prompt, first choice)

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.