Gemini 3.1 Pro Explained: Preview Status, Price and Specs
In short: Gemini 3.1 Pro is a Google AI model released on 19 February 2026, offered in the Gemini API as a preview model under the name gemini-3.1-pro-preview. It costs $2 per million input tokens and $12 per million output tokens for prompts up to 200,000 tokens, and $4 / $18 above that. It reads text, images, video, audio and PDFs, with a 1,048,576-token context window and up to 65,536 tokens of output. Google's newest Pro-class model is Gemini 4 Argon.
What is Gemini 3.1 Pro?
Gemini 3.1 Pro is a large language model in Google's Gemini family, available through the Gemini API, Google AI Studio and the Gemini app. Google's model page lists the model code as gemini-3.1-pro-preview, and a second endpoint, gemini-3.1-pro-preview-customtools, for developers building with a mix of bash and custom tools.
For how models like this work, see how LLMs work.
What does "preview" mean for Gemini 3.1 Pro?
A preview model is available to use but has not been declared a stable release. In Google's naming, preview models carry -preview in their code and can change or be replaced. Google's deprecations page lists gemini-3.1-pro-preview as released on 19 February 2026, with no shutdown date announced. The earlier gemini-3-pro-preview was shut down on 9 March 2026, with Gemini 3.1 Pro as its listed replacement.
How much does Gemini 3.1 Pro cost?
The price depends on prompt length: one rate up to 200,000 tokens, a higher rate above. From Google's pricing page:
| Per 1M tokens | Prompts up to 200k tokens | Prompts over 200k tokens |
|---|---|---|
| Input | $2 | $4.00 |
| Output | $12 | $18.00 |
| Cached input | $0.20 | $0.40 |
Output prices include thinking tokens. The Batch API halves input and output rates, and Gemini 3.1 Pro is not available on the Gemini API free tier.
What are Gemini 3.1 Pro's specs?
| Spec | Gemini 3.1 Pro |
|---|---|
| Developer | |
| Released | February 19, 2026 |
| Status | Preview |
| Model code | gemini-3.1-pro-preview |
| Input token limit | 1,048,576 |
| Output token limit | 65,536 |
| Input | Text, images, video, audio and PDF |
| Output | Text |
| Tools | Function calling, structured output, code execution, Google Search grounding |
| Licence | Proprietary - not downloadable |
How does Gemini 3.1 Pro relate to Gemini 4 Argon?
Gemini 4 Argon is Google's newer model, generally available in the Gemini API since 13 October 2026. It costs $2 input and $10 output per million tokens during its introductory period, with a 1.05M-token context window. Gemini 3.1 Pro remains available as a preview model. See Gemini 4 Argon explained.
How does Gemini 3.1 Pro rank?
On LMArena's text leaderboard Gemini 3.1 Pro has 1487 Elo (rank 20). Gemini 4 Argon has 1520 Elo (rank 2, high effort). The full ranked table is on our benchmarks page.
Words to know
- Preview model - a model available to use but not yet a stable release.
- Model code - the exact name used to call a model in the API.
- Token - a chunk of text or media; prices are per million.
- Input / output token limit - the most tokens a model takes in, and writes, per request.
- Tiered pricing - a higher rate once a prompt passes a size threshold.
- Batch API - sending many requests to be processed later at a discount.
Related guides
Frequently asked questions
Is Gemini 3.1 Pro still a preview?+
Yes. Google lists it as gemini-3.1-pro-preview, released on 19 February 2026, with no shutdown date announced.
How much does Gemini 3.1 Pro cost?+
For prompts up to 200,000 tokens it costs $2 per million input tokens and $12 per million output tokens. Above 200,000 tokens the rates rise to $4 and $18.
What is Gemini 3.1 Pro's context window?+
Its input token limit is 1,048,576 tokens and its output token limit is 65,536 tokens.
Can Gemini 3.1 Pro understand audio and video?+
Yes. Gemini 3.1 Pro accepts text, images, video, audio and PDF files as input, and returns text.
What replaced Gemini 3 Pro?+
Google shut down gemini-3-pro-preview on 9 March 2026 and listed gemini-3.1-pro-preview as its replacement.
Enjoyed this? Get more in your inbox.
Up to ten of the week's AI news stories, by email once a week.