Flower Labs unveiled its Endeavor 1.0 model, positioning it as a frontier‑class alternative to OpenAI’s GPT‑5.6 and Anthropic’s Claude Fable 5, touting local deployment and data sovereignty.
Flower Labs, a spinout from Cambridge University, has unveiled Endeavor 1.0, a large‑language model it describes as a “frontier‑class” alternative to OpenAI’s upcoming GPT‑5.6 and Anthropic’s Claude Fable 5.
A home‑grown model for data‑sovereign AI
Endeavor 1.0 is built to run on on‑premise hardware, allowing enterprises and public sector organisations to keep sensitive data within their own firewalls. The company says the model can be fine‑tuned on proprietary datasets without sending any information to external cloud services, a claim that resonates with growing concerns over data privacy in Europe.
According to Flower Labs, the model matches the performance of GPT‑5.6 and Claude Fable 5 on a suite of benchmark tests, including reasoning, code generation, and multilingual understanding. While exact scores are not disclosed, the firm highlights comparable results on the widely used Helix and Omega evaluation suites.
Technical highlights
- 7.2 billion parameters, optimized for low‑latency inference
- Hybrid transformer‑Mixture‑of‑Experts architecture for scalable compute
- Native support for on‑device quantisation down to 4‑bit precision
- Open‑source tooling for model inspection and safety alignment
The model also incorporates a safety layer trained on a curated dataset of ethically flagged content, aiming to reduce hallucinations and biased outputs. Flower Labs plans to release an API that can be self‑hosted, giving customers full control over usage limits and logging.
Market positioning and competition
By targeting organisations that require strict data governance, Endeavor 1.0 positions itself away from the public‑cloud‑centric offerings of OpenAI and Anthropic. Analysts note that while the model may not yet match the sheer scale of the biggest LLMs, its deployment flexibility could carve a niche in regulated sectors such as finance, healthcare, and government.
Flower Labs’ CEO, Dr. Amelia Hart, told Tech.eu that the company aims to “democratise access to cutting‑edge AI while respecting sovereignty,” and that the next roadmap includes a 15‑billion‑parameter version slated for early 2027.
Endeavor 1.0 gives organisations the power to run world‑class AI locally, without sacrificing performance or safety.
The launch comes as European regulators tighten rules around AI transparency and data handling, making locally hosted models an increasingly attractive proposition for compliance‑focused businesses.
Comments
No comments yet.