Gemini 3.1 Pro lands — Google's reasoning leap, and what it means for Australians (2026)

ℹ Editor's note, 23 September 2026: Google renamed NotebookLM to Gemini Notebook in July 2026. This is a dated record and keeps the name that was correct when it was written; the review carries the current name.

What changed

Google has released Gemini 3.1 Pro, an upgrade it describes as "a step forward in core reasoning." The headline number is a benchmark: on ARC-AGI-2 — a test of how well a model handles entirely new logic patterns it hasn't seen before — Gemini 3.1 Pro scores 77.1%, more than double the result of the Gemini 3 Pro model it replaces.

In plain terms, that points to a model that holds up better on genuinely hard, multi-step problems: synthesising messy data into a single dashboard, generating animated code-based visuals from a prompt, or building interactive 3D experiences.

The rollout is broad. Developers get it through the Gemini API (Google AI Studio, Gemini CLI, Antigravity and Android Studio in preview). Enterprises get it on Vertex AI and Gemini Enterprise. And everyday users get it in the Gemini app and NotebookLM, with higher usage limits for Google AI Pro and Ultra subscribers.

What it means for Australians

For most Australian Gemini users, this is a free upgrade you don't have to do anything to receive — Google rolls these models out globally through the app, and the reasoning gain shows up where the work is hardest. If you've used Gemini for a quick rewrite and found it fine but unremarkable, this is the kind of release that's worth a second look on your thornier tasks.

Google's standing local advantage holds: it runs Australian data centres in Sydney and Melbourne, which is part of why Gemini already rates well on Australian Fit in our reviews. Google's announcement doesn't break out Australian pricing or availability specifics for this model, so paid limits will follow whatever plan you're on.

Does it change our recommendation

It reinforces it. Google Gemini was already one of the strongest-value AI assistants for Australians, and a reasoning jump of this size widens its lead on complex work without asking you to pay more. If you're weighing it against ChatGPT or Perplexity, it's a good moment to retest on the tasks you actually care about.

We score every tool against the same five criteria, including Australian Fit. See our Methodology for how that works.

Sources

• Gemini 3.1 Pro announcement — Google — Primary, checked 19 February 2026. Confirms the 19 February release, the 77.1% ARC-AGI-2 score (more than double Gemini 3 Pro), and the API / Vertex AI / Gemini app / NotebookLM rollout.
• Wayback snapshot queued (SPN pending); OG card to render on promotion.