Gemini 3.8 Flash in NeuralSpace: chat, code, and API

Gemini 3.8 Flash is now available in NeuralSpace for three everyday workflows: chat, project work in Code, and public API calls. In Chat and Code you can choose reasoning effort, while the model accepts text, images, and function calls. That makes it useful for more than quick answers: it can stay with a multi-step task and use the available tools when needed.

Where to select it

Open Gemini 3.8 Flash chat to discuss a task, analyse text, or work from an image. For development, select it in Code, where it can use the project context and the available tools. Choose a lower reasoning effort for fast iterations or a higher one for a more careful pass.

Use it through the API

The public API model ID is gemini-3-8-flash at /v1/chat/completions. It uses the OpenAI Chat Completions format: send messages and, when needed, function definitions in tools. See the NeuralSpace API documentation for authentication and request examples.

A practical first run

1. State the outcome you need in a short first message.
2. Add code, an image, or project context only when it changes the answer.
3. Use low reasoning effort for quick iterations; raise it for difficult diagnosis.
4. When the task needs a project action, let the model call the relevant tool and review its result before proceeding.

The interface shows the cost before you run a request; API usage is calculated from actual input and output tokens.