MINA LABS
Start creating free
GitHub has made Gemini 3.8 Flash available in GitHub Copilot. The announcement, published in the GitHub Changelog, describes it as Google’s latest Flash model and points to early testing on complex terminal-based coding tasks.
GitHub says the model performed strongly in those tests and demonstrated rigorous reasoning. The announcement does not provide a broader benchmark table, detailed pricing for Copilot users, or a complete list of supported workflows. The useful takeaway is narrower: Gemini 3.8 Flash is now an option inside GitHub’s coding assistant.
Terminal-based work is often where coding tasks become less self-contained. A model may need to interpret a repository, follow several commands, understand tool output, and adjust its next step based on what happened. That is different from producing a short code snippet in response to a single prompt.
GitHub’s early testing is relevant because it focuses on this kind of work. If the model can reason through complex terminal tasks, it may be useful for projects where the answer depends on the state of a codebase rather than on one isolated file. That could include investigating a failing command, tracing a build problem, or planning a sequence of repository changes. The announcement does not claim that every task succeeds, so these should be treated as areas to evaluate rather than guaranteed capabilities.
The model’s availability in GitHub Copilot also gives developers a way to try it within an existing coding workflow. There is no need to treat the announcement as a separate model release that requires a new type of project. Its immediate significance is access: developers using GitHub Copilot can consider Gemini 3.8 Flash when choosing how to approach demanding coding work.
The word “Flash” also sets an expectation about the product category, but not a complete performance profile. The source highlights terminal reasoning, while leaving other questions open. It does not specify response latency, context limits, supported programming languages, or how Gemini 3.8 Flash compares with other models available through GitHub Copilot. Those details matter when selecting a model for production work and would need to be tested directly.
What it costs and whether it is on Mina Labs
Gemini 3.8 Flash is available on Mina Labs. The price for Gemini Omni Flash is 17 for the listed five-second generation duration.
That pricing is separate from whatever access terms apply inside GitHub Copilot. The Changelog announcement confirms availability in Copilot, but it does not state that Copilot access and Mina Labs usage share the same billing system. For teams comparing workflows, the practical distinction is simple: Copilot provides the model within GitHub’s development environment, while Mina Labs provides access here for making and evaluating generations.
What we would use it for
We would start with repository tasks that require several steps and clear verification. For example, we would ask Gemini 3.8 Flash to inspect a project, explain a failing terminal command, propose a small fix, and describe how to check whether the change worked. The point would not be to accept every suggested edit automatically. It would be to see how well the model maintains the thread from diagnosis through validation.
We would also test it on unfamiliar codebases. A useful trial would provide the model with a project structure and a concrete objective, then measure whether its instructions remain specific as the task develops. We would look for accurate references to files and commands, sensible handling of errors, and explanations that make it possible for a person to review the work.
For creative and technical teams, that makes Gemini 3.8 Flash worth evaluating when a task is more involved than a one-shot code request. The current announcement supports a focused claim: GitHub has added the model to Copilot, and its early testing found strong performance on complex terminal-based coding tasks. Mina Labs makes the model available here as well, so builders can test whether that strength fits their own projects.