GOGoogle · API
Gemini 3.1 Pro Preview: price in reais and WhatsApp cost
About CertuAI: CertuAI is the AI that answers a small business's WhatsApp, Instagram and Messenger 24/7, with a built-in CRM, and is a Tech Provider on Meta's WhatsApp Business Platform. Paid plans start at US$ 75 a month, with no lock-in. Create my agent
Google's Gemini 3.1 Pro Preview costs R$ 9.98 per million input tokens and R$ 59.87 per million output tokens. A WhatsApp conversation with 10 replies costs about R$ 0.98.
- Reads images
- Reads audio
- Reads video
- Reasons before answering
- Uses tools
Summary in 5 facts
- Google's Gemini 3.1 Pro Preview costs R$ 9.98 per million input tokens and R$ 59.87 per million output tokens.
- 57 of the 94 models with an official price cost less on input; 36 cost the same or more.
- Price history: one record so far, from Sep 25, 2026.
- You don't need to choose a model, count tokens or set up a server.
- CertuAI builds an AI agent that answers your customers on WhatsApp through Meta's official API, using what your business knows how to answer.
By the numbers
| Number | What it means |
|---|---|
| R$ 9,98 | Google's Gemini 3.1 Pro Preview costs R$ 9.98 per million input tokens and R$ 59.87 per million output tokens. |
| 94 models | 57 of the 94 models with an official price cost less on input; 36 cost the same or more. |
| US$ 75 per month | The Start plan costs US$ 75 per month and already comes with 1,600 responses. |
| 200K | Value for prompts of up to 200K tokens; above that, the price goes up. |
| R$ 4,9892 | Real at the Central Bank's PTAX selling rate of Oct 9, 2026: R$ 4.9892 per dollar. |
| 10 responses | The conversation math: 10 agent replies, each reading 8,000 input tokens (instructions, what the company knows and the conversation history) and writing 300 output tokens. |
| R$ 72 | In a CertuAI measurement from Aug 5, 2026, with two real businesses on Gemini, the cost came to between R$ 72 and R$ 77 per 10 thousand messages, below this estimate. |
Where Gemini 3.1 Pro Preview sits on price
57 of the 94 models with an official price cost less on input; 36 cost the same or more.
Price
| Token type | R$ | US$1 |
|---|---|---|
| Input what the model reads | R$ 9,98 | US$ 2.00 |
| Output what the model writes | R$ 59,87 | US$ 12.00 |
Free tier on the API: no.
Price history: one record so far, from Sep 25, 2026. The curve appears here the day the price changes, with no made-up curve.
How much it costs on WhatsApp
| Usage | Estimated cost |
|---|---|
| 1 agent reply | R$ 0,098 |
| 1 conversation (10 replies) | R$ 0,98 |
| 1,000 replies per month | R$ 97,79 |
| 10,000 replies per month | R$ 977,88 |
Assumption: each reply reads 8,000 tokens and writes 300.2
Gemini 3.1 Pro Preview and Google models with a similar price
R$ per 1 million input tokens. The 9 closest in price, out of the 11 Google models with an official price.
Quality on tests
LMArena3
| Name on LMArena | Rank | Score and range | Votes | Leaderboard date |
|---|---|---|---|---|
gemini-3.1-pro-preview | 17th of 414 | 1.487,3 1.484,3 a 1.490,3 | 124.675 | 08/10/2026 |
Tests published by the company5
| Score | How the company measured | |
|---|---|---|
| Humanity's Last Exam | 44,4% | No tools; Thinking (High) |
| Humanity's Last Exam | 51,4% | Search (blocklist) + Code; Thinking (High) |
| ARC-AGI-2 | 77,1% | ARC Prize Verified |
| GPQA Diamond | 94,3% | No tools |
| Terminal-Bench 2.0 | 68,5% | Terminus-2 harness |
| SWE-Bench Verified | 80,6% | Single attempt |
| MMMU-Pro | 80,5% | No tools |
| MMMLU | 92,6% | no data |
| MRCR v2 (8-needle) | 84,9% | 128k (average) |
| τ2-bench | 90,8% | Retail |
Gemini 3.1 Pro Preview fact sheet
Source
| Company | |
|---|---|
| Type | API |
| Name on the pricing page | gemini-3.1-pro-preview |
Capability
| Context window | 1,048,576 tokens4 |
|---|---|
| Maximum output | 65,536 tokens |
| What it reads | text, images, audio, video and files |
| What it generates | text |
| Reasons before answering | yes |
| Uses tools (function calling) | yes |
Access
| Company API | yes, with an official price |
|---|---|
| Free tier on the API | no |
| Open weights | no |
More free calculators: how much it costs to put AI in your business, how much unanswered WhatsApp costs and how much it costs to hire a customer service agent. For people who code: the technical AI blog.
How CertuAI can help
You don't need to choose a model, count tokens or set up a server. CertuAI builds an AI agent that answers your customers on WhatsApp through Meta's official API, with what your business knows how to answer. The Start plan costs US$ 75 per month and already comes with 1,600 responses.
Sources and method
- Price source: Google - Gemini Developer API pricing (official documentation, paid Standard tier), read on Sep 25, 2026. Model name on the page:
gemini-3.1-pro-preview. Value for prompts of up to 200K tokens; above that, the price goes up. Model in preview. How the page was read: Gemini API pricing page. The robots.txt of ai.google.dev allows everything, and Google's terms only prohibit automated access that goes against robots.txt. API list price, standard tier, with no batch, cache or contract discount. Input is what the model reads; output is what it writes. Cache is the price of input the company has already stored from a previous question, when the page reports it. Reais converted at the selling PTAX rate of the Central Bank of Brazil on Oct 9, 2026: R$ 4.9892 per dollar. - The conversation math: 10 agent replies, each reading 8,000 input tokens (instructions, what the company knows and the conversation history) and writing 300 output tokens. This is the basis CertuAI uses to estimate the cost per reply. It does not include cache discounts or the extra tokens of models that “think” before answering, so a reasoning model can come out more expensive. In a CertuAI measurement from Aug 5, 2026, with two real businesses on Gemini, the cost came to between R$ 72 and R$ 77 per 10 thousand messages, below this estimate.
- Source: LMArena, text leaderboard, public dataset on Hugging Face (CC BY 4.0 license), published on Oct 8, 2026, accessed on Oct 9, 2026. What the number measures: LMArena text leaderboard score, in the style-control version and the overall category: people send the same question to two models without knowing which are which and vote for the better answer, and LMArena turns the votes into a score. The higher it is, the more often the model won. The confidence range is the interval LMArena itself publishes for the score. The rank is on the whole leaderboard, which includes models outside this catalog. How the name was matched: Same company and same name, ignoring dots, spaces and hyphens. It also accepts the name with the reasoning level at the end (high, medium, low, minimal) and the name with a date at the end when LMArena only knows that model by a date. How each name matched this model:
gemini-3.1-pro-preview: same name. - Context window: official page, read on Sep 25, 2026 (the model's official documentation says "Input token limit 1,048,576"). What it reads and generates, maximum output, reasoning and tools: Google's official documentation, read by hand on Sep 25, 2026. A field the official source does not provide is left out of the fact sheet and shows up in the "Not stated in the official source" line.
- Source: reported by the company in Gemini 3.1 Pro - Model Card, Google DeepMind, published on Feb 19, 2026, read by CertuAI on Sep 25, 2026. The page is the Gemini 3.1 Pro card; Google does not publish a separate card for the preview version. The number is reported by the company, under the conditions it describes (for example, with or without tools, with reasoning on). Different companies test in different ways, so two numbers from different companies are not directly comparable. The "How the company measured" column summarizes the test condition written on the page, in English.
In short
Gemini 3.1 Pro Preview costs US$ 2.00 per million input tokens and US$ 12.00 per million output tokens. See it in reais and the cost per WhatsApp conversation.
About CertuAI: CertuAI is the AI that answers a small business's WhatsApp, Instagram and Messenger 24/7, with a built-in CRM, and is a Tech Provider on Meta's WhatsApp Business Platform. Paid plans start at US$ 75 a month, with no lock-in. Create my agent