# Gemini 3 Pro

> Google's first Gemini 3 flagship — a sparse mixture-of-experts model built for deep reasoning, agentic work, and multimodal understanding.

- Maker: Google
- Line: Gemini Pro
- Canonical page: https://ai-tldr.dev/models/gemini-3-pro/
- Updated: 2026-06-18

## In short

Gemini 3 Pro is the first flagship model in Google's Gemini 3 generation, released on November 18, 2025. It is the most capable tier of the Gemini 3 family at launch, positioned for complex reasoning, agentic workflows, and coding. It has a 1M-token context window. You can use it through Google AI / Gemini API and Google Vertex AI.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Context window | 1M (max output 64K) | [1] |
| Input price | $2.00 / 1M tokens (prompts ≤ 200K; $4.00 above 200K) | [2] |
| Output price | $12.00 / 1M tokens (prompts ≤ 200K; $18.00 above 200K) | [2] |
| Licence | Proprietary · API only | [1] |
| Released | 18 November 2025 | [3] |
| Knowledge cutoff | January 2025 | [1] |

## Sources

1. [Gemini 3 Pro model card (PDF)](https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Model-Card.pdf)
2. [Google — Build with Gemini 3 (developer guide)](https://blog.google/technology/developers/gemini-3-developers/)
3. [Google — Gemini 3 announcement](https://blog.google/products/gemini/gemini-3/)

## Overview

Gemini 3 Pro is the first flagship model in Google's Gemini 3 generation, released on November 18, 2025. It is the most capable tier of the Gemini 3 family at launch, positioned for complex reasoning, agentic workflows, and coding. Google describes Gemini 3 Pro as a sparse mixture-of-experts transformer with native multimodal support, trained from scratch rather than fine-tuned from the earlier Gemini 2.5 line.

The model accepts text, images, audio, video, and PDFs, with a 1 million-token context window and output up to roughly 64K tokens. Google calls Gemini 3 Pro its best model for multimodal understanding, and it topped the LMArena leaderboard at launch. Two new controls — a thinking_level parameter to dial reasoning effort and a media_resolution parameter for vision detail — let developers trade off latency, cost, and accuracy.

Gemini 3 Pro is available through the Gemini API in Google AI Studio, on Vertex AI, and inside the Gemini app. It launched at $2 per million input tokens and $12 per million output tokens for prompts of 200K tokens or less, with higher rates for longer prompts. It is a proprietary, closed-weights model. Google later shipped Gemini 3.1 Pro and Gemini 3.5 Pro as successors in this line.

## Benchmarks

- GPQA Diamond: 91.9% ([source](https://blog.google/products/gemini/gemini-3/))
- Humanity's Last Exam (no tools): 37.5% ([source](https://blog.google/products/gemini/gemini-3/))
- SWE-bench Verified: 76.2% ([source](https://blog.google/products/gemini/gemini-3/))
- Terminal-Bench 2.0: 54.2% ([source](https://blog.google/products/gemini/gemini-3/))
- MMMU-Pro: 81% ([source](https://blog.google/products/gemini/gemini-3/))
- Video-MMMU: 87.6% ([source](https://blog.google/products/gemini/gemini-3/))
- SimpleQA Verified: 72.1% ([source](https://blog.google/products/gemini/gemini-3/))
- MathArena Apex: 23.4% ([source](https://blog.google/products/gemini/gemini-3/))

## Pricing

- Input: $2.00 / 1M tokens (prompts ≤ 200K; $4.00 above 200K) per 1M tokens
- Output: $12.00 / 1M tokens (prompts ≤ 200K; $18.00 above 200K) per 1M tokens

Launch API pricing. Context caching and Batch API reduce costs further.

Source: https://blog.google/technology/developers/gemini-3-developers/

## Strengths

- State-of-the-art reasoning, scoring 91.9% on GPQA Diamond and 37.5% on Humanity's Last Exam (no tools)
- Strong agentic coding, with 76.2% on SWE-bench Verified and 54.2% on Terminal-Bench 2.0
- Best-in-class multimodal understanding across image, video, audio, and PDF inputs (81% MMMU-Pro, 87.6% Video-MMMU)
- 1M-token context window for reasoning over long documents, codebases, and mixed media
- Developer controls for reasoning depth (thinking_level) and vision detail (media_resolution) to tune cost and latency
- Topped the LMArena leaderboard at launch for general capability

## Best for

- Complex multi-step reasoning and analysis over long documents or research material
- Agentic coding tasks: bug fixing, repo-scale changes, and terminal-driven workflows
- Multimodal understanding combining text, images, video, audio, and PDFs
- Building autonomous agents and tool-using assistants via the Gemini API or Vertex AI
- Long-context tasks such as summarizing or querying entire codebases and large document sets
- Factual question answering with reduced hallucination (72.1% on SimpleQA Verified)

## How to access

- [Google AI / Gemini API](https://ai.google.dev/gemini-api/docs) — model id `gemini-3-pro-preview`
- [Google Vertex AI](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro) — model id `gemini-3-pro`

## Gemini Pro versions

- [Gemini 3.5 Pro](https://ai-tldr.dev/models/gemini-3-5-pro/) (19 May 2026)
- [Gemini 3.1 Pro](https://ai-tldr.dev/models/gemini-3-1-pro/) (19 February 2026)
- Gemini 3 Pro (18 November 2025) — this page
- [Gemini 2.5 Pro](https://ai-tldr.dev/models/gemini-2-5-pro/) (25 March 2025)
- [Gemini 2.0 Pro](https://ai-tldr.dev/models/gemini-2-0-pro/) (5 February 2025)
- [Gemini 1.5 Pro](https://ai-tldr.dev/models/gemini-1-5-pro/) (15 February 2024)
- [Gemini 1.0 Ultra](https://ai-tldr.dev/models/gemini-1-0-ultra/) (8 February 2024)
- [Gemini 1.0 Pro](https://ai-tldr.dev/models/gemini-1-0-pro/) (13 December 2023)

## FAQ

### When was Gemini 3 Pro released?

Google released Gemini 3 Pro on November 18, 2025, as the first flagship model in the Gemini 3 generation. It was followed by Gemini 3.1 Pro and Gemini 3.5 Pro.

### What is Gemini 3 Pro's context window?

Gemini 3 Pro has a 1 million-token context window and can produce up to roughly 64K tokens of output. It accepts text, images, audio, video, and PDF inputs.

### How much does the Gemini 3 Pro API cost?

At launch, Gemini 3 Pro cost $2 per million input tokens and $12 per million output tokens for prompts of 200K tokens or less. Prompts over 200K tokens are billed at $4 input and $18 output per million tokens. Context caching and the Batch API can lower costs further.

### Is Gemini 3 Pro open source?

No. Gemini 3 Pro is a proprietary, closed-weights model. It is available through the Gemini API in Google AI Studio, on Vertex AI, and in the Gemini app, but the weights are not publicly downloadable.

## Links

- [Google — Gemini 3 announcement](https://blog.google/products/gemini/gemini-3/)
- [Google — Build with Gemini 3 (developer guide)](https://blog.google/technology/developers/gemini-3-developers/)
- [Gemini 3 Pro model card (PDF)](https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Model-Card.pdf)
- [Gemini 3 Pro — Vertex AI documentation](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/3-pro)
- [Gemini API pricing](https://ai.google.dev/gemini-api/docs/pricing)
- [Google DeepMind — Gemini models](https://deepmind.google/models/gemini/)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/gemini-3-pro/