Google has officially unveiled Gemini 3 Flash, its newest fast and cost-efficient AI model, and is rolling it out as the default model in the Gemini app and Google Search’s AI Mode. Built on the Gemini 3 family released last month, the model delivers frontier-level intelligence with dramatically higher speed and lower cost, signaling Google’s most aggressive push yet in the AI race.

Gemini 3 Flash replaces Gemini 2.5 Flash globally, while users can still switch to Gemini 3 Pro for advanced math, reasoning, and coding tasks.

Why Gemini 3 Flash Matters

Despite being positioned as a speed-focused model, Gemini 3 Flash delivers strong results across major benchmarks — in some cases rivaling top-tier frontier models.

On Humanity’s Last Exam, which evaluates cross-domain expertise, Gemini 3 Flash scored 33.7% without tool use, compared with: Gemini 3 Pro: 37.5%, Gemini 2.5 Flash: 11%, GPT-5.2: 34.5%

The model also topped the MMMU-Pro multimodal reasoning benchmark, achieving an 81.2% score, the highest among competing models.

Google says these gains allow Gemini 3 Flash to match or exceed more expensive models while remaining fast enough for real-time and large-scale workflows.

Gemini 3 Flash delivers frontier intelligence across top benchmarks.

Image source: Google

Default in the Gemini App and AI Mode

Google is making Gemini 3 Flash the default model worldwide in the Gemini app and AI Mode in Search. The model is designed to better understand multimodal inputs, enabling users to: Upload videos and ask for feedback or analysis, Draw sketches and have the model identify them, Analyze audio recordings, Generate quizzes, summaries, tables, and visual answers

Users who need deeper reasoning can still select Gemini 3 Pro from the model picker.

Developers and Enterprises Are Already Using It

Google says companies including JetBrains, Figma, Cursor, Harvey, and Latitude are already using Gemini 3 Flash in production. The model is available through: Vertex AI, Gemini Enterprise, Gemini API (preview), Google Antigravity, Google’s new agentic coding platform

For developers, Gemini 3 Flash supports fast iteration, video analysis, data extraction, visual Q&A, and repeatable workflows with low latency.

Pricing and Efficiency

Gemini 3 Flash pricing: $0.50 per 1M input tokens, $3.00 per 1M output tokens

Although slightly more expensive than Gemini 2.5 Flash, Google says the model: Is 3× faster than Gemini 2.5 Pro, Uses 30% fewer tokens on average for reasoning tasks, Outperforms Gemini 2.5 Pro at a fraction of the cost

Google positions Flash as its primary workhorse model for high-volume AI workloads.

Frontier Intelligence at Scale

According to Google, Gemini 3 Flash delivers frontier-class performance on advanced benchmarks, including: GPQA Diamond: 90.4%, Humanity’s Last Exam: 33.7%, SWE-bench Verified: 78% (second only to GPT-5.2)

The model introduces advanced visual and spatial reasoning, along with code execution for tasks like zooming, counting, and editing visual inputs.

Image source: Google

The AI Race Intensifies

Google revealed that since releasing Gemini 3, it has processed over 1 trillion tokens per day across its API. The launch comes as competition with OpenAI intensifies, following the release of GPT-5.2 and a new image generation model.

While Google avoided directly addressing OpenAI, it said rapid releases across the industry are pushing all AI companies to innovate faster.

Google launches Gemini 3 Flash, makes it the default model in the Gemini app

Google has launched Gemini 3 Flash, making it the default model in the Gemini app and boosting AI Mode speed, efficiency, and real-time performance.