Gemini 3.1 Pro Model Card

This model card governs disclosure of Gemini 3.1 Pro, Google's flagship multimodal reasoning model published February 2026, and serves as a technical transparency document rather than a legally binding terms instrument; distribution is subject to separate terms including the Gemini API Additional Terms of Service and Google Cloud Platform Terms of Service depending on access channel. The document states that Gemini 3.1 Pro accepts inputs including text, images, audio, video, and code repositories with a 1M token context window, and produces text outputs up to 64K tokens, distributed across channels including the Gemini App, Vertex AI, Google AI Studio, Gemini API, Google Antigravity, Gemini Enterprise, and NotebookLM. The model card defers substantive disclosure of known limitations, acceptable usage policies, safety policies, risks and mitigations, training data, and hardware details to the Gemini 3 Pro model card rather than restating them, which creates a layered disclosure structure where material governance information is distributed across multiple documents. The document discloses frontier safety evaluations under Google's Frontier Safety Framework across five risk domains (CBRN, cyber, harmful manipulation, machine learning R&D, and misalignment), noting that Gemini 3.1 Pro reached the cyber alert threshold but did not reach the Critical Capability Level, and that ongoing mitigations are deployed in both the CBRN and cyber domains; this disclosure engages emerging regulatory frameworks including the EU AI Act's requirements for high-risk and general-purpose AI systems, and may also be relevant under voluntary AI safety commitments made by major AI developers to governments. Compliance teams should note that the document's reliance on cross-referenced disclosures in the Gemini 3 Pro model card means a complete governance assessment of Gemini 3.1 Pro requires review of that prior document, and that the cyber alert threshold disclosure and harmful manipulation evaluations may warrant evaluation under applicable AI governance frameworks and sector-specific regulations depending on deployment context.