Gemini 3.1 Pro is now available on Vertex AI (via the Gemini Enterprise Agent Platform), Gemini CLI, and Gemini Enterprise. The model brings a 1M token context window, native support for text, audio, images, video, PDFs, and code repositories, and up to a 15% improvement on complex problem-solving benchmarks relative to the best Gemini 3 Pro Preview runs — while also producing fewer output tokens, which reduces latency and cost at scale. For developers, Gemini 3.1 Pro is immediately accessible inside Gemini Code Assist in VS Code and IntelliJ, covering preview mode, agent mode, chat, and code generation. The release consolidates Google's top-tier model offering under the Gemini Enterprise Agent Platform umbrella rebranded from Vertex AI at Cloud Next '26, giving enterprise teams a single surface for both model access and agentic orchestration. Teams already building on Gemini 3 Pro Preview should evaluate the upgrade path given the performance and token-efficiency gains.