Base44 has launched its own AI model, Base 1, targeting the vibe-coding market dominated by competitors like Lovable, Replit, and Cursor. According to Business Insider, CEO Maor Shlomo stated that the San Francisco-headquartered startup trained its own large language model to reduce the AI-slop design that has become characteristic of vibe-coded products. The company, which is a subsidiary of website-building company Wix, aims to differentiate from competitors who use underlying frontier models like Anthropic's Opus 4.8, OpenAI's GPT-5.5, or Gemini 3.
Recent testing pitted Base 1 against Anthropic's Opus 4.8 to build the same website, with Base 1 demonstrating superior performance across multiple metrics. As reported by Business Insider, the test revealed that Base 1 was faster and used fewer credits to deliver a competitive product compared to the frontier model. CEO Shlomo emphasized that the model would produce faster results and cost users fewer credits than using frontier models, positioning it as a more cost-effective solution for businesses seeking AI-powered design capabilities.
The new model addresses a critical issue in the vibe-coding market where websites built with frontier models often share similar interfaces and design elements. According to Business Insider, design experts have noted that these sites look like an algorithmic Uniqlo or Ikea - functional but basic. Shlomo explained that Base 1, in its more advanced versions in the future, will create something that looked uniquely different every time, though he acknowledged that "It's going to take some effort, so we are not yet there." This differentiation strategy could give Base44 a competitive edge in the increasingly cost-conscious AI market.
Booking.com's chief business officer James Waters has revealed how the online travel giant is leveraging AI for strategic analysis and operational efficiency. According to Business Insider, Waters uses AI models including Claude, Gemini, and ChatGPT for opposition research, analyzing how competitors tackle major strategic problems that would otherwise take days to research manually. Recently, he used Anthropic's Claude to analyze customer reviews across platforms, demonstrating how AI can synthesize broad strategic landscapes in fraction of traditional time.
Three major AI developers have launched new models promising enhanced cost efficiency as businesses face mounting AI spending pressures. According to reports from Bloomberg, OpenAI announced its GPT-5.6 model, designed to complete more work while using significantly fewer tokens - a unit of data processed by AI models. SpaceXAI's Grok 4.5 is positioned as having twice the token efficiency compared to comparable models from other firms, while Meta is making its Muse Spark 1.1 pricing "very attractive," as CEO Mark Zuckerberg told Bloomberg.
Major companies are implementing comprehensive AI spending tracking systems as token costs escalate. According to Business Insider, Uber execs recently said the company blew through Claude Code budget for 2026, highlighting the challenge of managing AI spending. Booking.com tracks its total investment in tokens and people across value generated, with Waters stating that "the problem is, if those two things don't run together." The company measures token spending against value generation and tracks where tokens are spent to determine when cheaper models can be used for simpler tasks.
