DeepMind’s Gemini 2.0 Flash now rolls out to all developers via Google AI Studio and Vertex AI, replacing the earlier experimental release. The model offers a 1‑million‑token context window and multimodal input, enabling high‑volume, low‑latency tasks. With image‑generation and text‑to‑speech on the horizon, the update positions Gemini as a versatile workhorse for production workloads and scalability.

Alongside Flash, DeepMind released Gemini 2.0 Pro, an experimental model tailored for coding and complex prompts. With a 2‑million‑token context and tool‑calling abilities—including Google Search and code execution—the Pro variant promises sharper reasoning and broader knowledge. Available in AI Studio, Vertex AI, and the Gemini app for Advanced users, it reflects the company’s push toward higher‑performance, specialized models today.

DeepMind also unveiled Gemini 2.0 Flash‑Lite, a cost‑efficient model matching Flash’s speed while improving quality. Maintaining the 1‑million‑token window, Flash‑Lite excels on benchmarks and can caption 40,000 photos for under a dollar in paid tiers. Safety upgrades include reinforcement learning critiques and automated red‑team testing to curb prompt injection risks, underscoring DeepMind’s focus on secure deployment for enterprise and developers.