Back to events
Model releaseLimited global significanceConfirmed confidence

Google DeepMind releases Gemini 3 Flash

The lightweight variant of Google's third-generation architecture, optimized for time-to-first-byte and multi-task throughput under high concurrency.

Event details

Gemini 3 Flash is the lightweight variant of Google's third-generation architecture, focused on time-to-first-byte latency and multi-task throughput under high concurrency. Google DeepMind released the closed-source model on December 17, 2025; the parameter count is undisclosed and it is an autoregressive Transformer. The primary source is Google DeepMind's official model card.

Why it matters

Rated on the impact-v1 six-axis scale: technical 2/5, market 2/5, ecosystem 2/5, policy & social 1/5, scope 3/5, durability 2/5.

45/100Global significance score. Regional effects are recorded only when the evidence supports a meaningful difference.