Last week Apple made an announcement that it was ditching OpenAI, and partnering Google Gemini to power Siri.

Today Super Engineer is announcing that we too have ditched OpenAI in favour of Gemini to power our core LLM stack.

We started this re-architecture back in November, when it became clear that OpenAI's models (we were using GPT5 at the time) were:

• Less technically accurate

• Slower (by up to x4 times)

For us, answer accuracy / limiting hallucinations is core to the success of Super Engineer - as we don't want our users blowing stuff up. 

Other than accuracy, we also chose Gemini because of Google's wider tech stack, in particular Android XR - which we're building Super Spex (our AI Glasses product) around.

Importantly, the success of Super Engineer isn't just about AI - it's about the wider product design, and the ability for users to make the most of AI - hence why considering the wider ecosystem around our core tech partners is important, and why we're equally excited about Android XR's potential as we are with Gemini.

Thank you @SundarPichai @Gemini @AndroidXR and teams & we look forward to giving technicians & service engineers around the world Super Powers together