Google has released Gemini 3 Flash, completing its Gemini 3 family and signaling a major push toward faster, more capable, and more efficient artificial intelligence across its products.
The new model follows the earlier launch of Gemini 3 Pro and is designed to serve as Google’s high-speed, general-purpose option. Gemini 3 Flash is being deployed immediately to the Gemini app and web interface, where it will become the default model for most users. It is also rolling out to Google Search through AI Mode and is available to developers via the Gemini API, Vertex AI, AI Studio, and Antigravity.
According to Google, Gemini 3 Flash significantly improves on the previous 2.5 Flash model, narrowing the gap to the more powerful Pro version. Benchmark testing shows gains in academic reasoning, coding ability, and general knowledge accuracy. In advanced subject testing, the model’s performance reportedly approaches that of Gemini 3 Pro while maintaining much higher speed and lower computational cost.
Coding performance is one of the most notable upgrades. Gemini 3 Flash shows large improvements on industry-standard software engineering benchmarks, reflecting Google’s broader effort to make advanced coding assistance available without requiring its highest-tier model. The model has also reduced error rates on general knowledge questions, a longstanding challenge for fast, lightweight AI systems.
Efficiency is central to the Flash release. Google says the model runs workloads up to three times faster than earlier Pro models while retaining capabilities such as interactive simulations and multimodal outputs. Pricing reflects its middle-ground position: Gemini 3 Flash is more expensive than the older 2.5 Flash but far cheaper than Gemini 3 Pro, making it a more attractive option for developers balancing performance and cost.
For consumers, the rollout simplifies model selection inside the Gemini app. Both the “Fast” and “Thinking” settings now use Gemini 3 Flash, while Gemini 3 Pro remains available as a premium option. This change is intended to reduce confusion caused by the rapid pace of Google’s AI releases and to ensure that free users benefit from the latest improvements by default.
Source: ars