The New StackWednesday · May 20, 2026FREE

Google’s Gemini 3.5 Flash beats the frontier models

geminigoogleai-modelsllm

At its I/O developer conference on May 19, 2026, Google announced Gemini 3.5 Flash, a new AI model that it says beats frontier models on several benchmarks. The model is optimized for speed and cost, with pricing set at $0.10 per million input tokens and $0.40 per million output tokens. It is available immediately in preview via Google AI Studio and Vertex AI. Gemini 3.5 Flash supports a 1 million token context window and multimodal inputs including text, images, audio, and video. Google claims it outperforms GPT-4o and Claude 3.5 Sonnet on reasoning, coding, and math tasks. The model is part of Google's strategy to offer a lighter, faster alternative to its larger Gemini Ultra models, targeting developers building real-time applications.

// why it matters

Developers gain a cheaper, faster alternative to top-tier models without sacrificing performance.

Sources

Primary · The New Stack
▸ Read original at thenewstack.io

Like this? Get the next digest.

Google’s Gemini 3.5 Flash beats the frontier models — aigest.dev