The cheap new AI model taking aim at OpenAI and Anthropic

Type Safe AI, founded by former OpenAI researcher Diogo Almeida, has released its model "Jev" as a cheaper alternative to large language models like ChatGPT and Claude for certain tasks. The start-up, recently valued at $200mn, has drawn multibillion-dollar valuation offers of $10bn or more from potential investors. Jev, aimed almost exclusively at software developers, has gone viral on social media, with its launch video on X generating 40mn views in less than a week.

The company is betting that routine tasks currently handed to expensive general-purpose models can be performed by cheaper, more specialised systems, potentially pressuring the business models of frontier AI companies such as OpenAI and Anthropic. Jev does not produce sentences or images but promises fast, cheap decisions within software applications, such as approving or blocking requests and routing support tickets. James Hardiman, general partner at DCVC, which led Type Safe's most recent funding round, said the start-up was already profitable given the drastically lower computing costs for its product, making its model "orders of magnitude" cheaper to use than leading LLMs.

Type Safe AI，由前OpenAI研究员Diogo Almeida创立，已发布其模型"Jev"，作为ChatGPT和Claude等大型语言模型的更便宜替代方案，适用于某些任务。该创业公司近期估值2亿美元，已从潜在投资者那里获得了100亿美元或更高的估值报价。Jev主要面向软件开发人员，其在X平台上的发布视频