Upgrade to Gemini 3 Deep Think for research
TL;DR
A major upgrade to the specialized reasoning mode for science and engineering is now available to Google AI Ultra subscribers in the Gemini App.
## What changed Google released a major upgrade to Gemini 3 Deep Think on February 12, 2026. The specialized reasoning mode now combines deep scientific knowledge with practical engineering utility to tackle real problems in science, research, and engineering.
The update is available only to Google AI Ultra subscribers inside the Gemini App. Users select Deep Think in the prompt bar, submit a task, and receive a response after a few minutes of processing.
## Why it matters This move targets users who need reliable performance on hard technical work instead of fast general answers. It puts pressure on other frontier models that still struggle with multi-step science and engineering tasks. Google is betting that a subset of paying users will value slower, deeper reasoning enough to stay on the Ultra tier.
## How to use it Subscribe to Google AI Ultra through the Gemini App or one.google.com. Open the app, choose Deep Think from the prompt bar options, and enter your query. Wait for the notification that the response is ready.
## Watch for Clear confirmation appears when independent tests show consistent gains on science and engineering benchmarks. The bet weakens if most users stick to faster modes due to the wait time. The next logical step is integration of Deep Think into the new Mac desktop app.
Harsh’s take
For a solo operator this upgrade gives access to stronger technical reasoning without extra hires, yet it requires the highest subscription tier and forces you to plan around multi-minute delays. The real trade-off is deciding whether occasional depth on research tasks justifies paying more than the Pro plan while accepting slower iteration than Flash or Pro.
Most Vibe Builders will still default to faster models for daily work. Test Deep Think on one actual engineering or research problem this week and measure output quality against your current setup before committing to the cost.
by Harsh Desai
About Gemini
View the full Gemini page →All Gemini updatesGo deeper
More AI news
- Daily RoundupGLM 5.3 and Vercel agent tools land, Gemini SAT practice, plus fresh models on Hugging Face
Google and Vercel released new agent and model features while Hugging Face and Replicate added trending models; OpenAI tightened safety after agent issues.
- Daily RoundupVercel-Origin beta and GPT-5.6 Sol discount, plus agent tools for builders
Vercel rolled out Origin repository deployments and a month-long GPT-5.6 Sol discount on AI Gateway while Product Hunt featured new agent layers and Hugging Face highlighted an uncensored Qwen model, with industry moves around data use and automation exits.
- Weekly DigestHermes Agent v0.20 voice and A2A protocol, OpenClaw mobile updates, and multi-agent tools to test now
Hermes Agent rolled out v0.20 releases with voice streaming, A2A interoperability, desktop artifacts, and self-recovery tools while OpenClaw added Android switching, iOS caching, and plugin provenance checks across the week.