Gemini 4 Neon is presented as a high-capability model for complex software engineering, knowledge work, and scientific workflows, emphasizing fast completion of tasks that may require processing large amounts of information. It accepts text, images, video, audio, and PDFs, and produces text. Its unified context window is 1,048,576 tokens, covering input and output together. Standard synchronous requests allow up to 65,536 output tokens; a long-decode option can extend output to 1,048,576 tokens. The listed stable model identifier is gemini-4-neon, with a latest update dated November 2026.
The model supports agentic video understanding, code execution, computer use, file search, function calling, structured outputs, and grounding through Google Search and Google Maps. It also offers URL context and configurable thinking levels from low through xhigh. Context caching is available only implicitly, not through explicit caching. Audio generation, image generation, and the Live API are not supported, despite audio and image being accepted as inputs. For deployment, it supports the Batch API and flex and priority inference options. The documentation points readers to separate guidance for feature changes, pricing, migration, and comparisons with other Gemini models, so those details are not specified in this overview.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.