Google released Gemini 3.5 Flash on May 19, 2026, adding a lightweight proprietary multimodal model to the Gemini 3.5 family. The release is distinct from the earlier Gemini 3.1 Flash Lite and Gemini 3.1 Pro launches and represents the first publicly available 3.5-generation model. The move continues Google's pattern of shipping tiered model variants to cover a wider range of cost and latency targets. A lightweight 3.5-generation model positions Google to compete with fast, low-cost inference tiers offered by OpenAI and Anthropic, and gives developers building latency-sensitive applications a newer baseline to benchmark against.