Gemini 4 Neon now available on AI Gateway
Gemini 4 Neon is available through AI Gateway with tool calling, structured output, multimodal input, and four reasoning levels. Builders can test it in coding-agent workflows using the new model ID.
**Gemini 4 Neon** is available on AI Gateway under `google/gemini-4-neon`. It accepts text, images, video, and documents, and supports a **1M-token context window**, tool calling, structured outputs, and Google Search and Maps grounding.
Builders can try it in coding-agent workflows that need tool use or multi-step tasks. Its reasoning settings are **low, medium, high, and xhigh**.
**Gemini 4 Neon** is available on AI Gateway under `google/gemini-4-neon`. It accepts text, images, video, and documents, and supports a **1M-token context window**, tool calling, structured outputs, and Google Search and Maps grounding. Builders can try it in coding-agent workflows that need tool use or multi-step tasks. Its reasoning settings are **low, medium, high, and xhigh**. Vercel describes its output quality as close to larger Gemini 4 models, with faster responses than the earlier lighter Gemini 4 model. The material provides no benchmark results to quantify those comparisons.
This adds a multimodal Gemini 4 route with video and document input, tool calling, structured output, and Google Search and Maps grounding alongside its 1M-token context. Vercel’s quality and speed comparisons to larger Gemini models remain unquantified, so the listing expands the options to evaluate without establishing a preferred route.