Google launches Gemini 3.1 Pro: What’s new in Google’s latest AI model and who gets access first
&imwidth=600&imheight=450&format=webp&quality=medium "Google launches Gemini 3.1 Pro: What’s new in Google’s latest AI model and who gets access first")
Google launches Gemini 3.1 Pro Photograph: (AI generated)
Google has launched Gemini 3.1 Pro, upgrading its flagship AI with stronger reasoning, coding and scientific problem-solving skills. The model scores higher on key benchmarks and is rolling out in preview across Gemini app, NotebookLM and developer platforms.
Google has released Gemini 3.1 Pro, the latest upgrade to its flagship AI model, focusing on stronger reasoning and better performance in complex tasks.
As per the Google’s official blog, the new model is designed for situations where “a simple answer is not enough.” It is rolling out in preview across the Gemini app, NotebookLM, and developer platforms including Google AI Studio and Vertex AI.
The company says this update brings meaningful improvements in reasoning, coding, scientific understanding and agent-based workflows.
What is Gemini 3.1 Pro?
Gemini 3.1 Pro is an upgraded version of Google’s Gemini 3 Pro model. Unlike earlier updates that used a “.5” naming pattern, this is Google’s first “.1” version.
The company has indicated that this is a focused intelligence upgrade rather than a broad feature expansion.
Trending Stories
&im=FitAndFill=(700,400))
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
)
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
The model integrates the reasoning system introduced in Gemini 3 Deep Think and makes it available to a wider group of users.
Google states that Gemini 3.1 Pro is built to handle:
Large and messy datasets
Multi-step technical questions
Advanced scientific reasoning
Complex coding tasks
Key Features: What’s New in Gemini 3.1 Pro?
1. Stronger reasoning for complex tasks
Google says the model can synthesise large datasets into a single clear output. This means it can combine multiple inputs and deliver structured responses across layered problems.
2. Animated SVG generation from text
One new capability is generating animated SVG files directly from a text prompt.
Unlike video files, SVGs are built in code, so they remain sharp at any size and use smaller file sizes.
3. Better coding and agent workflows
Gemini 3.1 Pro improves performance in “agentic coding” tasks, where AI handles multiple steps autonomously.
4. Improved scientific reasoning
The model shows strong results in scientific benchmarks, particularly in chemistry and physics-related problem solving.
Benchmark Scores: How Gemini 3.1 Pro Performs
Google shared several benchmark results in its announcement:
77.1 per cent on ARC-AGI-2 (abstract reasoning benchmark)
This is more than double the previous Gemini 3 Pro score.
94.3 per cent on GPQA Diamond (scientific knowledge benchmark)
80.6 per cent on SWE-Bench Verified (coding tasks benchmark)
85.9 per cent on BrowseComp (agentic search benchmark)
2887 Elo on LiveCodeBench Pro (competitive coding performance)
According to Google, these scores show measurable improvements in reasoning, coding and multi-step problem solving compared to earlier versions.
Gemini 3.1 Pro vs GPT and Claude
While Google did not directly name competitors in the blog, the benchmark comparisons place Gemini 3.1 Pro against leading frontier models from OpenAI and Anthropic.
On ARC-AGI-2, the 77.1per cent score represents a significant improvement in abstract reasoning. Competitive coding Elo also places the model among top-performing AI systems in that category.
However, cross-platform comparisons depend on independent testing conditions and real-world usage.
Who Can Access Gemini 3.1 Pro?
Gemini 3.1 Pro is currently launching in preview mode.
Consumers
Available in the Gemini app
Available in NotebookLM
Higher usage limits for Google AI Pro and Ultra subscribers
Developers
Access via:
Google AI Studio
Gemini API
Vertex AI
Antigravity
Gemini Enterprise
Android Studio
Google says general availability is expected soon after performance validation, especially around agent-based workflows.
Why This Update Matters
The focus on reasoning suggests that AI development is shifting from simple prompt responses to structured problem solving.
Gemini 3.1 Pro is aimed at users who require:
Technical problem solving
Research-level reasoning
Coding automation
Structured data interpretation
With improved benchmarks and expanded access, Google is positioning Gemini 3.1 Pro as a stronger competitor in the advanced AI model space.
