Gemini 3 Flash on streaming video
Gemini 3 Flash · Google DeepMind · Multimodal · Closed API
Gemini 3 Flash is Google DeepMind’s high-throughput Gemini tier, released in February 2026 with a 1M-token context window at streaming-friendly prices. On Overshoot it is available as a passthrough model, reachable through the same stream publishing and OpenAI-compatible chat-completions request format used across the API.
Gemini 3 Flash is available through the Overshoot API as a passthrough model (gemini-3-flash-preview, status: ready, verified against the live model catalog on 2026-07-14) - requests are forwarded to the upstream provider rather than running on Overshoot's hosted infrastructure.
- Developer
- Google DeepMind
- Parameters
- Undisclosed
- Context window
- 1M tokens
- License
- Proprietary
- Released
- Feb 2026
- Inputs
- Text, images, video frames
- Overshoot availability
- API passthroughas of 2026-07-14
What Gemini 3 Flash is good at
Gemini 3 Flash delivers quality close to Google’s frontier tier at a fraction of the cost and latency, which makes it a strong default for high-volume visual workloads. It handles everyday scene understanding, object detection, and document reading with the accuracy expected of a current-generation frontier lab model.
Its 1M-token context window is unusually large for a fast tier, so an application can keep long conversation history, tool outputs, and many sampled video frames in a single request without needing to prune earlier context.
- Frontier-adjacent visual accuracy at streaming-friendly cost
- A 1M-token context window for extended sessions
- Solid document and scene understanding at high request volume
Running Gemini 3 Flash on Overshoot
Publish a camera or screen share over WebRTC through LiveKit to open a Stream, then send a chat-completions request whose image_url or video_url is an ovs:// reference to the latest frame, an exact timestamp, or a recent segment drawn from the 600-second retention window.
As a passthrough model, Gemini 3 Flash reaches the API through the same OpenAI-compatible endpoint used across Overshoot: the platform manages stream lifecycle, frame selection, and SSE streaming transport, then forwards the resolved request to Google. The current passthrough id in the live catalog is gemini-3-flash-preview, as returned by GET /v1beta/models.
Gemini 3 Flash versus Gemini 3 Pro
Flash trades some reasoning depth for much higher throughput and lower cost compared with Gemini 3 Pro, which targets the hardest long-video and document understanding tasks. Applications that query frames frequently or at scale typically default to Flash and reserve Pro for requests that need its extra reasoning ceiling.
Frequently asked questions
Can Gemini 3 Flash analyze live video?
Yes. Through Overshoot, Gemini 3 Flash answers questions about a live WebRTC stream. Your application references the latest frame or a recent segment with an ovs:// URL in a standard chat-completions request, and Overshoot forwards the resolved request to Google.
Is Gemini 3 Flash open source?
No. Gemini 3 Flash is a proprietary Google DeepMind model available only through Google’s API. Overshoot reaches it as a passthrough model, using the same key management as its hosted open-weight models.
How fast is Gemini 3 Flash compared to Gemini 3 Pro?
Flash is built for high throughput and lower latency than Pro, at some cost to peak reasoning depth. Both are passthrough models on Overshoot, so overall response time depends on Google’s serving stack, but Flash is consistently the quicker tier.
What context window does Gemini 3 Flash support?
Gemini 3 Flash supports a 1M-token context window, large enough to hold extensive conversation history and many sampled video frames in a single request, unusual for a fast, low-cost tier.