Google DeepMind
Build with Gemini Try Gemini

Gemini

Our most intelligent AI models

Chat with Gemini Build with Gemini

Models

Gemini 2.5 Flash-Lite is now ready for scaled production use

Gemini 2.5 models are capable of reasoning through their thoughts before responding, resulting in enhanced performance and improved accuracy.

Model family

Gemini 2.5 builds on the best of Gemini — with native multimodality and a long context window.

Create and edit images with Gemini 2.5 Flash Image

Generate, transform and edit images with simple text prompts, or combine multiple images to create something new. All in Gemini.

Learn more

Hands-on with Gemini 2.5

See how Gemini 2.5 uses its reasoning capabilities to create interactive simulations and do advanced coding.

Make an interactive animation

See how Gemini 2.5 Pro uses its reasoning capabilities to create an interactive animation of “cosmic fish” with a simple prompt.

Create your own dinosaur game

Watch Gemini 2.5 Pro create an endless runner game, using executable code from a single line prompt.

Code a fractal visualization

See how Gemini 2.5 Pro creates a simulation of intricate fractal patterns to explore a Mandelbrot set.

Plot interactive economic data

Watch Gemini 2.5 Pro use its reasoning capabilities to create an interactive bubble chart to visualize economic and health indicators over time.

Animate complex behavior

See how Gemini 2.5 Pro creates an interactive Javascript animation of colorful boids inside a spinning hexagon.

Code particle simulations

Watch Gemini 2.5 Pro use its reasoning capabilities to create an interactive simulation of a reflection nebula.

Adaptive and budgeted thinking

Adaptive controls and adjustable thinking budgets allow you to balance performance and cost.

  • Calibrated

    The model explores diverse thinking strategies, leading to more accurate and relevant outputs.

  • Controllable

    Developers have fine-grained control over the model's thinking process, allowing them to manage resource usage.

  • Adaptive

    When no thinking budget is set, the model assesses the complexity of a task and calibrates the amount of thinking accordingly.

Gemini 2.5 Deep Think

An enhanced reasoning mode that uses cutting edge research techniques in parallel thinking and reinforcement learning to significantly improve Gemini’s ability to solve complex problems.

Deep Think can better help tackle problems that require creativity, strategic planning, and making improvements step-by-step.

  • Iterative development and design

    We’ve seen impressive results on tasks that require building something by making small changes over time.

  • Aiding scientific and mathematical discovery

    By reasoning through complex problems, Deep Think can act as a powerful tool for researchers.

  • Algorithmic development and code

    Deep Think excels at tough coding problems where problem formulation and careful consideration of tradeoffs and time complexity is paramount.

Performance

Gemini 2.5 is state-of-the-art across a wide range of benchmarks.

View 2.5 tech report

Benchmarks

In addition to its strong performance on academic benchmarks, Gemini 2.5 tops the popular coding leaderboard WebDev Arena.

Benchmark
Gemini 2.5
Flash-Lite Non-thinking
Gemini 2.5
Flash-Lite Thinking
Gemini 2.5
Flash Non-thinking
Gemini 2.5
Flash Thinking
View 2.5 Flash
Gemini 2.5
Pro Thinking
View 2.5 Pro
Input price
$/1M tokens
(no caching) 		$0.10 $0.10 $0.30 $0.30 $1.25
 $2.50 > 200k tokens
Output price
$/1M tokens $0.40 $0.40 $2.50 $2.50 $10.00
 $15.00 > 200k tokens
Reasoning & knowledge Humanity's Last Exam (no tools)
5.1% 6.9% 8.4% 11.0% 21.6%
Science GPQA diamond
64.6% 66.7% 78.3% 82.8% 86.4%
Mathematics AIME 2025
49.8% 63.1% 61.6% 72.0% 88.0%
Code generation LiveCodeBench (UI: 1/1/2025-5/1/2025)
33.7% 34.3% 41.1% 55.4% 69.0%
Code editing Aider Polyglot
26.7%
27.1%
44.0%
56.7%
82.2%
Agentic coding SWE-bench Verified
single attempt 31.6% 27.6% 50.0% 48.9% 59.6%
   
multiple attempts 42.6% 44.9% 60.0% 60.3% 67.2%
Factuality SimpleQA
10.7% 13.0% 25.8% 26.9% 54.0%
Factuality FACTS grounding
84.1% 86.8% 83.4% 85.3% 87.8%
Visual reasoning MMMU
72.9% 72.9% 76.9% 79.7% 82.0%
Image understanding Vibe-Eval (Reka)
51.3% 57.5% 66.2% 65.4% 67.2%
Long context MRCR v2 (8-needle)
128k (average) 16.6% 30.6% 34.1% 54.3% 58.0%
   
1M (pointwise) 4.1% 5.4% 16.8% 21.0% 16.4%
Multilingual performance Global MMLU (Lite)
81.1% 84.5% 85.8% 88.4% 89.2%

Methodology

Gemini results: All Gemini scores are pass @1."Single attempt" settings allow no majority voting or parallel test-time compute; "multiple attempts" settings allow test-time selection of the candidate answer. They are all run with the AI Studio API with default sampling settings. To reduce variance, we average over multiple trials for smaller benchmarks. Aider Polyglot score is the pass rate average of 3 trials. Vibe-Eval results are reported using Gemini as a judge. Google's scaffolding for "multiple attempts" for SWE-Bench includes drawing multiple trajectories and re-scoring them using model's own judgement. For Aider results differ from the official leaderboard due to a difference in the settings used for evaluation (non-default).

Result sources: Where provider numbers are not available we report numbers from leaderboards reporting results on these benchmarks: Humanity's Last Exam results are sourced from https://agi.safe.ai/ and https://scale.com/leaderboard/humanitys_last_exam, LiveCodeBench results are from https://livecodebench.github.io/leaderboard.html (1/1/2025 - 5/1/2025 in the UI), Aider Polyglot numbers come from https://aider.chat/docs/leaderboards/. FACTS come from https://www.kaggle.com/benchmarks/google/facts-grounding. For MRCR v2 which is not publically available yet we include 128k results as a cumulative score to ensure they can be comparable with other models and a pointwise value for 1M context window to show the capability of the model at full length. The methodology has changed in this table vs previously published results for MRCR v2 as we have decided to focus on a harder, 8-needle version of the benchmark going forward.

Input and output price reflects text, image and video modalities.

Building responsibly in the agentic era

As we develop these new technologies, we recognize the responsibility it entails, and aim to prioritize safety and security in all our efforts.

Learn more

For developers

Gemini’s advanced thinking, native multimodality and massive context window empowers developers to build next-generation experiences.

Start building

Developer ecosystem

Build with cutting-edge generative AI models and tools to make AI helpful for everyone.

Google AI Studio

Build with the latest models from Google DeepMind

Gemini API

Easily integrate Google’s most capable AI model to your apps

Accessing our latest AI models

We want developers to gain access to our models as quickly as possible. We’re making these available through Google AI Studio.

Sign in to Google AI Studio

Get the latest updates

Sign up for news on the latest innovations from Google DeepMind.

Gemini Flash-Lite

General availability

Gemini 2.5 Flash-Lite

Best for high volume, cost efficient tasks

Build with Gemini

Introducing 2.5 Flash-Lite, a thinking model for those looking for low cost and latency.

Upgrade to Gemini 2.5

2.5 Flash-Lite excels at high-volume, latency-sensitive tasks like translation and classification.

  • Thinking, enabled

    Experience improved reasoning and output quality with thinking mode and thinking budgets.

  • Superior latency

    Benefit from faster response times.

  • Tool use

    Utilize key Gemini 2.5 features including tool uses like Search and code execution.

  • Cost-efficient

    2.5 Flash-Lite is our most cost-efficient 2.5 model yet.

Hands-on with 2.5 Flash-Lite

Build a dynamic UI

See how Gemini 2.5 Flash-Lite writes the code for a UI and its contents based solely on the context of what appears in the previous screen - all in the time it takes to click a button.

Watch

Turn PDFs into interactive web apps

See how Gemini 2.5 Flash-Lite built a research prototype that can instantly transform large PDF files into interactive web apps - making it easier to summarize and understand dense information.

Watch

Benchmarks

2.5 Flash-Lite has all-round, significantly higher performance than 2.0 Flash-Lite on coding, math, science, reasoning and multimodal benchmarks.

Benchmark
Gemini 2.0
Flash
Gemini 2.5
Flash-Lite Non-thinking
Gemini 2.5
Flash-Lite Thinking
Reasoning & knowledge Humanity's Last Exam (no tools)
5.1%* 5.1% 6.9%
Science GPQA diamond
65.2% 64.6% 66.7%
Mathematics AIME 2025
29.7% 49.8% 63.1%
Code generation LiveCodeBench (UI: 1/1/2025-5/1/2025)
29.1% 33.7% 34.3%
Code editing Aider Polyglot
21.3% 26.7% 27.1%
Agentic coding SWE-bench Verified
single attempt 21.4% 31.6% 27.6%
   
multiple attempts 34.2% 42.6% 44.9%
Factuality SimpleQA
29.9% 10.7% 13.0%
Factuality FACTS grounding
84.6% 84.1% 86.8%
Visual reasoning MMMU
69.3% 72.9% 72.9%
Image understanding Vibe-Eval (Reka)
55.4% 51.3% 57.5%
Long context MRCR v2 (8-needle)
128k (average) 19.0% 16.6% 30.6%
   
1M (pointwise) 5.3% 4.1% 5.4%
Multilingual performance Global MMLU (Lite)
83.4% 81.1% 84.5%

Methodology

Gemini results: All Gemini scores are pass @1."Single attempt" settings allow no majority voting or parallel test-time compute; "multiple attempts" settings allow test-time selection of the candidate answer. They are all run with the AI Studio API with default sampling settings. To reduce variance, we average over multiple trials for smaller benchmarks. Aider Polyglot score is the pass rate average of 3 trials. Vibe-Eval results are reported using Gemini as a judge. Google's scaffolding for "multiple attempts" for SWE-Bench includes drawing multiple trajectories and re-scoring them using model's own judgement. For Aider results differ from the official leaderboard due to a difference in the settings used for evaluation (non-default).

Result sources: Where provider numbers are not available we report numbers from leaderboards reporting results on these benchmarks: Humanity's Last Exam results are sourced from https://agi.safe.ai/ and https://scale.com/leaderboard/humanitys_last_exam, LiveCodeBench results are from https://livecodebench.github.io/leaderboard.html (1/1/2025 - 5/1/2025 in the UI), Aider Polyglot numbers come from https://aider.chat/docs/leaderboards/. FACTS come from https://www.kaggle.com/benchmarks/google/facts-grounding. For MRCR v2 which is not publically available yet we include 128k results as a cumulative score to ensure they can be comparable with other models and a pointwise value for 1M context window to show the capability of the model at full length. The methodology has changed in this table vs previously published results for MRCR v2 as we have decided to focus on a harder, 8-needle version of the benchmark going forward.

* these results are on an earlier HLE dataset, obtained from https://scale.com/leaderboard/humanitys_last_exam_preview

Model information

2.5 Flash-Lite
Model deployment status General availability
Supported data types for input Text, Image, Video, Audio, PDF
Supported data types for output Text
Supported # tokens for input 1M
Supported # tokens for output 64k
Knowledge cutoff January 2025
Tool use Search as a tool
Code execution
Best for High volume, low-cost and low latency tasks
Availability Google AI Studio
Gemini API
Vertex AI