Google: Gemini 3.1 Pro
Gemini 3.1 Pro is the Pro-tier model in Google's Gemini 3 series, available in the Gemini API as gemini-3.1-pro-preview. According to the Gemini API release notes, Google released it on February 19, 2026 as its latest iteration in the Gemini 3 family, together with a gemini-3.1-pro-preview-customtools endpoint. On March 9, 2026 the gemini-3-pro-preview alias was repointed to gemini-3.1-pro-preview. The id keeps the -preview suffix, and Google's deprecations page lists no shutdown date for it (checked 2026-10-29).
Model Specs
The model page lists these specs:
| Spec | Gemini 3.1 Pro Preview |
|---|---|
| Model id | gemini-3.1-pro-preview |
| Input token limit | 1,048,576 |
| Output token limit | 65,536 |
| Inputs | Text, image, video, audio and PDF |
| Output | Text only |
| Supported | Caching, code execution, function calling, Google Maps and Google Search grounding, structured outputs, thinking, URL context, file search (AI Studio only) |
| Not supported | Audio generation, image generation, the Live API |
| Latest update | February 2026 |
The page does not list a knowledge cutoff.
Key Features
- Long multimodal input: about one million tokens of context across text, images, video, audio and PDFs, with text output.
- Thinking and tools: thinking, function calling, structured outputs, code execution and grounding with Google Search or Maps.
- Custom-tools variant:
gemini-3.1-pro-preview-customtoolsis a separate endpoint for workflows that mix bash with custom tools such asview_fileorsearch_code. Google notes quality may fluctuate on tasks that do not use them. - Consumption options: Batch API, Flex inference and Priority inference are listed on the model page.
Pricing
Google's pricing page, checked 2026-10-29, shows Standard rates per 1M tokens for Gemini 3.1 Pro Preview:
| Component | Prompts up to 200k tokens | Prompts over 200k tokens |
|---|---|---|
| Input | $2.00 | $4.00 |
| Output (including thinking tokens) | $12.00 | $18.00 |
| Context caching | $0.20 | $0.40 |
Caching storage is listed at $4.50 per 1M tokens per hour. The Batch tier is listed at roughly half of the Standard input and output prices, and Priority is priced higher. Confirm the live page before budgeting.
Use Cases
- Analysing long documents, codebases or video with one request.
- Agent workflows that need function calling plus grounding.
- Structured extraction from mixed media such as PDFs and audio.
A limitation: it is a preview id with text-only output, so image generation, audio output and the Live API need other models. Google's models page now describes Gemini 4 Argon as its flagship reasoning model, so check whether a newer Pro-tier model fits better.
Getting Started
- Create an API key in Google AI Studio and call
gemini-3.1-pro-previewfrom the Gemini API. - Pick a thinking setting and test on your own long-context examples.
- Use the custom-tools endpoint only if your agent mixes bash with custom tools.
- Estimate cost with the 200k-token pricing split in mind.
FAQ
What is the Gemini 3.1 Pro API id?
gemini-3.1-pro-preview, plus gemini-3.1-pro-preview-customtools for custom-tool workflows.
What is its context window?
Google lists an input limit of 1,048,576 tokens and an output limit of 65,536 tokens.
How much does Gemini 3.1 Pro cost?
Standard pricing is $2 input and $12 output per 1M tokens for prompts up to 200k tokens, and $4 and $18 above that.
Did it replace Gemini 3 Pro?
Google's March 9, 2026 release note says the gemini-3-pro-preview alias now points to gemini-3.1-pro-preview. See Gemini 3 Pro.
Alternatives
- Gemini 4 Argon: Google's newer frontier model.
- Gemini 3 Pro: the earlier Gemini 3 Pro preview.
- Gemini 3.6 Flash: a cheaper Flash tier model.
- Claude Opus 5.5: Anthropic's model for long-running agentic work.
Conclusion
Gemini 3.1 Pro is a capable long-context, multimodal Pro model that is still on Google's live models page as a preview. Check the official model page for current limits, and browse more in the models category.
Sources
- Websitehttps://ai.google.dev/gemini-api/docs/models
- Social Mediahttps://x.com/GoogleAI
Comments
No comments yet. Be the first to comment!
Related Tools
Related Insights
Point Codex at any model with magpie
Install magpie, add a local or signed-in provider, then switch Codex to that model. Rechecked against official magpie docs on 2026-10-06.

After Code Stopped Being Scarce (Part II): From AI Writing Code to AI Taking Part in Delivery
A year after leaving my previous employer, I document the delivery workflow for our financial business that we built around Claude Code and Opus 5.5: XaC, progressive rules, independent review and testing by Grok Bot, PR watch, and our actual delivery over the past three months.

Anthropic Subagent: The Multi-Agent Architecture Revolution
Deep dive into Anthropic multi-agent architecture design. Learn how Subagents break through context window limitations, achieve 90% performance improvements, and real-world applications in Claude Code.