← back
A year of Gemini progress + what comes next — Logan Kilpatrick, Google DeepMind
Takeaway
Gemini's last year was organizational consolidation plus omnimodal capability; the next year shifts scaffolding into the model itself and pushes proactive, agentic behavior.
Summary
- Logan Kilpatrick announces a final Gemini 2.5 Pro update with SOTA on AIDER and HLE; framed as turning point for Gemini.
- DeepMind sees 50x year-over-year increase in AI inference served, partly thanks to merging research, model and product teams under DeepMind.
- Roadmap: omnimodal mainline model (native audio/TTS in Gemini Live, VO video, diffusion experiments for high tokens/sec), and 'agentic by default' models that absorb scaffolding into reasoning.
- Gemini positioned as the unifying thread across all Google products, with proactivity called out as the next product frontier.
geminifoundation-modelsgoogle-deepmind
Original description
Over the last year, Google and Gemini models have shown rapid progress across all dimensions (model, product, etc). Let's highlight all the work that has happened, how we got the worlds best models, and where we are going next (across both the model landscape and out AI products). About Logan Kilpatrick Logan leads product for Google AI Studio and works on the Gemini API. Before Google, Logan led developer relations at OpenAI. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter