GitHub adds Microsoft's MAI-Code-1.1-Flash model to Copilot with vision support and lower list pricing

GitHub has started rolling out Microsoft's MAI-Code-1.1-Flash model inside GitHub Copilot, adding another low-cost coding model option across the company's developer assistant surfaces. The August 11 changelog says the new small-tier model builds on MAI-Code-1-Flash with native vision support for image understanding, plus claimed gains in coding quality, instruction following, tool use and performance.

The pricing change is the clearest operational hook for teams watching AI coding costs. GitHub says serving and model-efficiency improvements have enabled a 73% lower list price than the earlier MAI-Code-1-Flash model. For annual GitHub Copilot subscribers that remain on premium request accounting, the model is charged at a 0.25x premium request multiplier, meaning it is positioned for lightweight or high-volume coding workflows where a top-end model may be unnecessary.

Where developers can use it

According to GitHub, MAI-Code-1.1-Flash is being made available in a broad set of Copilot surfaces: Copilot CLI, the Copilot cloud agent, the standalone GitHub Copilot app, Copilot Chat on GitHub, Visual Studio Code, Visual Studio, GitHub Mobile, JetBrains IDEs, Eclipse and Xcode. Copilot Free and Student users can receive the model through automatic model selection. Paid individual, Business and Enterprise plans can manually select it where the model picker is available.

Managed organizations will still need to take an extra step. GitHub says Copilot Business and Copilot Enterprise administrators must enable the MAI-Code-1.1-Flash policy in Copilot settings, and that policy is off by default. That makes the launch less of an automatic switch for enterprise developers and more of a new option for teams willing to approve another model in their Copilot estate.

Why it matters

The update shows Microsoft and GitHub continuing to segment coding assistants by cost, latency and task type. Vision support could make a lower-cost model useful for practical developer tasks that mix code with images, such as reading UI screenshots, interpreting error dialogs, or comparing a rendered page with a design reference. The broader strategic point is that Copilot is becoming less dependent on a single default model and more like a routed model platform, where administrators and users choose between capability and cost.

The main caveat is that the performance and pricing claims are GitHub and Microsoft's own. Developers considering it for production workflows should test it against their repositories, extension versions and governance policies before assuming the lower price translates into better overall value.

Sources

• github.blog

Cover photo by Daniil Komov on Pexels, used under the Pexels License.