Google launches Gemini 3 Flash: Faster, cheaper AI with Pro-level intelligence
Google Gemini 3 Flash, a new AI model, is designed to deliver frontier-level intelligence at much faster speeds and lower cost. The model combines Gemini 3’s advanced reasoning with Flash-level efficiency and is now rolling out globally across Google products, developer platforms, and more.
Google announced the launch of Gemini 3 Flash, which is an extension of its latest Gemini 3 AI model line. The model has been developed to be fast, efficient, and inexpensive, in spite of possessing deep reasoning and multimodal capabilities.
As Google describes, Gemini 3 Flash brings next-generation AI within the reach of users, developers, and businesses.
What sets Gemini 3 Flash apart?
The Gemini 3 Flash integrates the brain of the Gemini 3 Pro and the flash performance and affordability of the other models.
Key highlights:
- Faster responses with lower delay
- Advanced reasoning and problem-solving
- Strong performance in multimodal tasks like text, images, video and audio
- Designed for high-frequency and real-time workflows
Google asserts that this model has particular ability in such tasks as agentic workflows.
Performance and Benchmark Results
Although Gemini 3 Flash is faster and more economical, it provides frontiersman-level performance on advanced tests.
Notable Scores:
- GPQA Diamond: 90.4 per cent
- Humanity’s Last Exam: 33.7 per cent (without tools)
- MMMU Pro: 81.2 per cent (comparable to Gemini 3 Pro)
Google asserts the performance of the newly introduced Google Gemini 3 Flash is better than Google Gemini 2.5 Pro with the use of 30% fewer tokens.
Built for speed and lower cost
- Speed is the biggest strength of Gemini 3 Flash.
- Up to 3x faster than Gemini 2.5 Pro
- Uses fewer tokens for everyday tasks
- Optimises “thinking time” based on task complexity
Pricing:
- USD 0.50 per 1M input tokens
- USD 3 per 1M output tokens
- Audio input: USD 1 per 1M tokens
This makes it one of Google’s most cost-efficient advanced AI models so far.
Big boost for developers
The target users of Gemini 3 Flash are developers who require rapid and smart AI assistance for coding and app development purposes.
Developer Benefits:
- Coding and Reasoning Skills Performance
- 78 per cent score on SWE-bench Verified, outperforming Gemini 3 Pro
- Suitable for agentic coding, interactive apps, and production systems
In addition to this, the model is capable of supporting intricate use cases such as video analysis and Q&A assistance as well as AI gaming assistants.
Where is Gemini 3 Flash available?
Starting from today, Gemini 3 Flash will be rolling out worldwide across multiple platforms:
- The Gemini App and AI Mode on Google Search
- Gemini API in Google AI Studio and Gemini CLI
- Google Antigravity (Agentic Development Platform)
- Vertex AI and Gemini Enterprise for businesses
and Gemini Enterprise capabilities for enterprises Already involved in this technology are large corporations such as JetBrains, Bridgewater Associates, and Figma.
Gemini 3 Flash is a major relief to Indian developers, startups and corporations, as it allows them to have cheaper and far more advanced AI technology in their applications, tools, and services.