Flower Labs

AI · Federated Learning · Enterprise AI

Launch

Cambridge spinout Flower Labs launches Endeavor to rival OpenAI, Anthropic

Local deployment means enterprises keep control even if a US provider abruptly restricts access, as happened to UK Claude users in June.

• Flower Labs launched Endeavor 1.0, a frontier-class generalist AI model for reasoning, coding, and long-horizon agent tasks, offered as a managed API service or fully self-hosted inside a customer's own infrastructure.
• Flower Labs is a Cambridge University spinout founded in 2023 by Nicholas Lane, Daniel Beutel, and Taner Topal, built on its open-source Flower federated-learning framework that trains models without moving data off customer systems.
• Endeavor scored 92.0 on GPQA, 98.2 on HumanEval, 99.9 on AIME 2026, and 94.1 on IFEval, results Flower Labs says put it on par with OpenAI's GPT-5.6 Sol and Anthropic's Claude Fable 5.
• The release arrives four months after Lizzy, Flower's 7B sovereign model for UK use cases, and is initially limited to a select group of partner organizations while the company adds compute capacity.
• Flower Labs has raised over £23 million since its 2023 Y Combinator debut and counts JPMorgan, IBM's Red Hat, and the NHS among more than 2,500 organizations using its models and infrastructure.
• The launch follows a June episode in which the US government and Anthropic temporarily restricted UK access to Claude Fable 5 over national security concerns, prompting Downing Street to seek an exemption.
• Endeavor tests whether federated-learning infrastructure can support a frontier-model business, offering Europe a credible alternative to renting intelligence from US API providers as the UK pushes a sovereign AI compute agenda.