An anonymous model called Ox Alpha app<\/a>eared on OpenRouter last week, free to use with a million-token context window, and developers have been impressed. OpenRouter\u2019s own listing says prompts and completions are retained by the unidentified provider.<\/p>\n A model that nobody will take credit for is being tested across the industry.<\/a> Ox Alpha appeared on OpenRouter last Thursday as a stealth release from an anonymous third-party provider, free to use, with a context window of just over a million tokens.<\/p>\n The scale on offer is not modest. The open-source agent OpenCode said the model would be free for a week with near unlimited usage, and that its provider had capacity for 100 trillion tokens a day.<\/p>\n Developers rate it. Strip<\/a>e\u2019s chief executive Patrick Collison tried it and called it \u201cvery impressive,<\/em>\u201d and it is positioned for coding, long-horizon agent work and production use.<\/p>\n The guessing game<\/a> has been inconclusive. The leading theory points to Z.ai, which previously tested GLM-5 anonymously under another name, while a competing analysis of its tokenizer suggests Microsoft\u2019s MAI family instead, in a market already reshaped by free Chinese models.<\/p>\n By the weekend the confidence had drained out of every theory. The AI analyst Andrew Curran wrote that people seemed \u201cless sure of anything<\/em>\u201d than they had been the night before.<\/p>\n The more useful detail is not the identity but the terms. OpenRouter\u2019s own listing states that prompts and completions \u201care retained by the provider and are not used for training.<\/em>\u201d<\/p>\n Read that again with a work project in mind. Whatever you send goes to a company that has not said who it is, and it keeps it.<\/p>\n For European businesses that is not an intrigue, it is a blocker. Data protection law requires a contract with a named processor and an assessment of where data goes, neither of which is possible when the counterparty is anonymous.<\/p>\n The timing sharpens it. The AI Act\u2019s transparency obligations took effect on 2 August, with penalties reaching \u20ac15mn or 3% of global turnover, in a regime built on knowing which provider is responsible for what.<\/p>\n None of this makes the model bad. Open-weight releases have closed the capability gap faster than the safety one, and a stealth launch is a legitimate way to benchmark a model before announcing it.<\/p>\n But free has a price here, and it is information. Somebody is paying for 100 trillion tokens a day of inference, and until they say who they are, the sensible European position is to test Ox Alpha with nothing that matters.<\/p>\n<\/p><\/div>\n If you liked the article, do not forget to share it with your friends. Follow us on\u00a0Google News<\/a><\/span>\u00a0too, click on the star and choose us from your favorites.<\/span><\/strong><\/p><\/blockquote>\n If you want to read more like this article, you can visit our Technology category.<\/a><\/span><\/strong><\/p>\n<\/blockquote>\n\n
![](\"https:\/\/media.thenextweb.com\/2026\/07\/big-tech-2-trillion-ai-spending-commitments-bloomberg.avif\")
<\/p>\n