All 416 tools and workflows the models named

How to read this

• Exa, Brave and Built-in count the models that recommended the product with that search route: out of 11, 11 and 4. A dot means no available answer named it; it is not a rate over questions. Qwen + Brave had nine read answers, Mistral + Exa 16, and MiniMax + Brave 17; other combinations had 18.
• Pain points come from one follow-up reply that linked the vendor’s own page, the one with the most filled pain-point fields, favoring the most common cited URL in a tie. Every answer was followed by the same question: what are the three main pain points each platform says it solves, according to its own website? These are the platforms’ claims as a model reported them, not our verdict.
• Follow-ups citing them is, for each route, how many follow-up replies that covered the product gave pain points with a link to the vendor, out of all that covered it. The rest said the claim couldn’t be verified, or got no reply. On Brave, follow-ups from Claude Opus 5.5, Claude Sonnet 5 and Gemini 3.8 Flash, and most from GLM 5.3, came back empty because of a fault on our side, so those missing replies cannot support a comparison of evidence quality between routes.
• Names and links were extracted by GPT-5.6 Luna. We reviewed names and cited URLs to merge clear aliases, keeping separate tools on the same domain distinct. Counts are deduplicated per product and response. We didn’t independently verify every identity or link. The alias decisions andcoverage gaps are available with the full analysis.

Read the post for what the overlap between combinations says, or try Run Any Model with a question of your own.