← Models

Gemini 3 Flash

Google · Gemini 3

Google's older Gemini 3 preview Flash route, now superseded by stable Gemini 3.5 Flash for most new fast-model work.

Part of Gemini Flash family · Other versions: Gemini 3.5 Live Translate , Gemini 3.5 Flash , Gemini 3.1 Flash-Lite , Gemini 3.1 Flash TTS Preview , Gemini 3.1 Flash Live Preview , Gemini 2.5 Flash Live Preview , Gemini 2.5 Flash-Lite , Gemini 2.5 Flash
Type
multimodal
Context
1M tokens
Max Output
66K tokens
Status
preview
Input
$0.5/1M tok
Output
$3/1M tok
API Access
Yes
License
proprietary
multimodal fast preview assistant tool-use long-context
Released December 2025 · Updated June 8, 2026

Overview

Freshness note: Model capabilities, limits, and pricing can change quickly. This profile is a point-in-time snapshot last verified on June 8, 2026.

Gemini 3 Flash is Google’s older preview-tier fast model in the Gemini API catalog, published as gemini-3-flash-preview. It originally gave teams a way to evaluate the Gemini 3 fast-model direction above the stable 2.5 Flash line. As of Google’s May 2026 Gemini 3.5 launch, new production-oriented fast/agentic work should usually start with Gemini 3.5 Flash, which is listed as stable.

This entry remains useful for teams that already evaluated or integrated the preview model and need to understand how it fits in the current Flash family.

Capabilities

Gemini 3 Flash is meant for the same broad class of work that made Flash useful in the first place: responsive assistants, extraction flows, multimodal chat, and agentic workloads that need good price-performance rather than maximum reasoning depth.

The current distinction is lifecycle and tool coverage. Gemini 3 Flash remains preview, while Gemini 3.5 Flash is the newer stable model for sustained frontier performance on agentic and coding tasks. Google’s docs still list Computer Use support on Gemini 3 Flash Preview; Gemini 3.5 Flash does not currently support Computer Use.

Technical Details

Google’s current Gemini API model catalog lists Gemini 3 Flash with:

  • Model ID: gemini-3-flash-preview
  • 1,048,576 token context window
  • 65,536 max output tokens
  • multimodal input support across text, images, audio, video, and files
  • thinking_level support instead of the older thinking_budget style
  • Computer Use Preview support for browser-control workloads

Google’s current model catalog lists Gemini 3 Flash as preview and Gemini 3.5 Flash as stable. The deprecation table lists no shutdown date for gemini-3-flash-preview, but it names gemini-3.5-flash as the recommended replacement. Treat Gemini 3 Flash as an evaluation, compatibility, or Computer Use route rather than the default new-build choice.

Pricing & Access

Google’s current Gemini API pricing still lists Gemini 3 Flash at:

  • Input: $0.50 per 1M text, image, or video tokens
  • Input: $1.00 per 1M audio tokens
  • Output: $3.00 per 1M tokens
  • Batch and Flex: 0.25text/image/videoinput,0.25 text/image/video input, 0.50 audio input, and $1.50 output per 1M tokens
  • Priority: 0.90text/image/videoinput,0.90 text/image/video input, 1.80 audio input, and $5.40 output per 1M tokens

Grounding with Search and Maps shares Google’s Gemini 3 allowance before per-query billing. Access is through Gemini API and Google AI Studio / Vertex AI surfaces where preview models are enabled.

Best Use Cases

Use Gemini 3 Flash for compatibility with existing preview-model tests, regression comparisons against the newer Gemini 3.5 Flash route, and controlled evaluation where preview behavior is acceptable.

For most new high-speed Gemini model selection, start with Gemini 3.5 Flash instead.

Comparisons

  • Gemini 3.5 Flash (Google): Newer stable fast/agentic route and the better default for new work.
  • Gemini 2.5 Flash (Google): Older stable production route for compatibility and cost comparisons.
  • Gemini 3.1 Flash-Lite (Google): Stable low-cost variant with a lower quality ceiling on difficult tasks.
  • GPT (OpenAI): Another fast production-model family choice, with the main tradeoff usually being platform strategy and multimodal tooling.