Gemini 3.1 Pro is a natively multimodal AI model from Google DeepMind designed for advanced reasoning, coding, document analysis, and agentic tasks. It is currently available through the Gemini API as gemini-3.1-pro-preview and has Preview status.
What is Gemini 3.1 Pro?
Gemini 3.1 Pro is an advanced model in the Gemini 3 series. It accepts text, images, video, audio, and PDF files, while producing text output. Its input context limit is 1,048,576 tokens and its maximum output is 65,536 tokens.
What can Gemini 3.1 Pro do?
The model combines multimodal understanding with reasoning and coding. Its official API documentation lists support for function calling, code execution, structured outputs, thinking, URL context, and grounding with Google Search. These capabilities make it suitable for long-document analysis, software development, research workflows, and AI agents.
What are its limitations?
Gemini 3.1 Pro remains a Preview model, so its behavior, features, or availability may change before a stable release. Its output is text only. According to the official model page, it does not support image generation, audio generation, or the Live API.
How much does Gemini 3.1 Pro cost?
On the paid tier, prompts up to 200,000 tokens cost $2 per million input tokens, $0.20 per million cached input tokens, and $12 per million output tokens. For prompts above 200,000 tokens, the rates rise to $4 for input, $0.40 for cached input, and $18 for output.
How does it perform on benchmarks?
Google DeepMind reports a score of 44.4% on Humanity’s Last Exam without tools and 77.1% on ARC-AGI-2. Benchmark results depend on the published test setup and do not guarantee the same performance in every real-world application.
Who is Gemini 3.1 Pro for?
It is aimed at developers, researchers, and teams that need long context, multimodal inputs, tool use, or complex reasoning. Teams building production systems should assess the operational risk of depending on a Preview model.
In short
Gemini 3.1 Pro combines a large context window with multimodal input, advanced reasoning, coding, and tool support. Its Preview status and higher pricing for prompts above 200,000 tokens are important limitations to consider.