Gemini 3 Flash

Gemini 3 Flash is Google's most balanced model, offering frontier-level reasoning capabilities at $0.50/$3 per million tokens—approximately 4x cheaper than Gemini 3 Pro while maintaining comparable intelligence for most tasks.

Gemini 3 Flash supports four thinking levels: minimal (near-zero latency), low, medium, and high—giving developers granular control over the reasoning depth vs. speed tradeoff that Gemini 3 Pro doesn't offer.

Yes, Gemini 3 Flash (gemini-3-flash-preview) has a free tier in the Gemini API, unlike Gemini 3 Pro which currently requires paid usage for API access.

Thought Signatures are encrypted representations of the model's internal reasoning that must be circulated back in multi-turn conversations—required even at minimal thinking level for Gemini 3 Flash to maintain reasoning context and enable function calling.

Yes, Gemini 3 Flash uniquely supports combining structured outputs (JSON schema) with built-in tools like Google Search, URL Context, and Code Execution in the same request—enabling grounded, type-safe responses.

The media_resolution parameter controls token usage per image/video frame: low (280 tokens), medium (560), high (1120), or ultra_high for images. For video, low and medium are both capped at 70 tokens per frame to optimize context usage.

Gemini 3 Flash supports Google Search, File Search, Code Execution, URL Context, and standard function calling. However, Google Maps grounding and Computer Use are not yet supported in Gemini 3 models.