Gemini 3.5 Pro

Google DeepMind · May 2026

activeClosedmixture of expertsmultimodalAPI Available
Context Window2M tokens

Description

Google's premium reasoning flagship, unveiled at I/O 2026 with a limited rollout. Represents Google's most capable model for complex coding, analysis, and multi-step reasoning tasks. General availability expected by end of June 2026.

Key Innovations

Reasoning
ReasoningStructured step-by-step problem solving, often using chain-of-thought or tree-of-thought approaches.
Multimodal
MultimodalProcessing multiple types of input (text, images, audio, video) in a single model.
Long Context
Long ContextAbility to process very long inputs (100K+ tokens), enabling analysis of entire codebases or books.

Family Tree