Mercury vs Brex
One of twelve models named Mercury first on the direct prompt; two named Brex. Both were named by all twelve models and Mercury carries 43 labels and Brex 24, so the shares are not directly comparable.
Named in seven categories this edition.
Named in twelve categories this edition.
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; every quote names the model and the prompt it came from. Both figures come from the business banking page.
By framing
MercuryFirst choices, of twelve modelsBrex
Across every category in the September 2026 Edition, Mercury and Brex were named in the same answer ninety-six times, of the 175 answers naming Mercury and the 539 naming Brex. In those answers Brex took the first choice nine times and Mercury twenty-two.
Every model, every framing
| Model | Direct | Paraphrase | Comparative | Budget-constrained | Scale-constrained | Negative |
|---|---|---|---|---|---|---|
| Claude Haiku 4.5 | | | | |||
| GPT-5.4 mini | | |||||
| Gemini 3.5 Flash | | | | | | |
| Perplexity Sonar | | | | | ||
| Grok 4.1 Fast | | | | | | |
| Mistral Small | | |||||
| DeepSeek V4 Flash | | | | | | |
| Llama 4 Maverick | | |||||
| Qwen 3.7 Flash | | | | |||
| Kimi K2 | | | | | ||
| GLM 4.7 FlashX | | | ||||
| MiniMax M2.5 | | |
Mercury Brex first choice named as an alternative argued againstblank: not namedEach cell is one answer, Mercury on the left and Brex on the right.The direct prompt
Mercury first, Brex an alternative
1 of 12 modelsBrex was named in the answer but not as the choice, or not at all.Brex first, Mercury an alternative
2 of 12 modelsMercury was named in the answer but not as the choice, or not at all.Neither was the first choice, one was named
4 of 12 modelsThe answer put something else first and named one of the two as an alternative.Neither was named
5 of 12 modelsThe answer made no first choice from these two in this category.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.
By buyer segment
Mercury33%#1 of 11Brex4%#6 of 11Brex18%#2 of 11Mercury9%#4 of 11What the models said about Mercury
Every negative label with a quote, up to three, then the highest-weighted positives, up to three. Six of seven in this category shown.
“Frequent account freezes/closure, slow support, funds held 60+ days | 15.4 negative reviews/month (4x industry average)” DeepSeek V4 Flash · negative prompt · hard negative
“neobanks (such as Mercury, Bluevine, or Novo) rely heavily on automated fraud-detection algorithms” Gemini 3.5 Flash · negative prompt · soft negative
“does not support physical branches or cash deposits” Perplexity Sonar · budget prompt · soft negative
“I'd usually recommend Mercury if your priority is a modern, startup-friendly banking platform” GPT-5.4 mini · paraphrase prompt · first choice
“For most mid-market B2B companies today, Mercury or Chase are the safest starting points” Kimi K2 · direct prompt · first choice
“Mercury, which is considered the best overall digital business banking platform for startups” Llama 4 Maverick · paraphrase prompt · first choice
What the models said about Brex
Every negative label with a quote, up to three, then the highest-weighted positives, up to three. Three of three in this category shown.
“Brex abruptly decided to stop servicing traditional small businesses, giving tens of thousands ... just 60 days to move their funds” Gemini 3.5 Flash · negative prompt · hard negative
“Brex is best if you run a mid-market company and want a corporate card plus treasury and vault services in one platform.” Claude Haiku 4.5 · direct prompt · first choice
“Brex offers both banking and corporate card services designed for venture-backed startups and mid-market companies” Claude Haiku 4.5 · 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.