.svg){kind=spacer}
Details
- Google AI Developers announced Gemini 3.1 Pro is now live in Augment Code's model picker, enabling devs to use its advanced reasoning and debugging directly in workspaces.
- The integration targets high-performance, lower-cost AI for faster, more efficient large-scale codebase changes.
- Gemini 3.1 Pro excels in complex tasks like agentic coding, tool use, and handling 1M token contexts for entire codebases, PDFs, images, video, and audio.
- Key upgrades include improved reasoning on benchmarks, Authentic Vision to reduce hallucinations in image analysis, and capabilities for interactive 3D simulations and real-time data visualizations.
- Available via Gemini API, Vertex AI, AI Studio; rolling out to Pro/Ultra users in Gemini app and NotebookLM, with preview endpoints like gemini-3.1-pro-preview-customtools for custom tools.
- Supports multi-step tasks, function calling, structured outputs, and code execution, optimized for software engineering and long-context analysis.
Impact
This integration pressures rivals like Anthropic's Claude and OpenAI's GPT-4o by embedding Google's frontier reasoning model directly into Augment Code, streamlining dev workflows with 1M token context for full codebase handling at lower cost. It accelerates adoption in agentic coding, where Gemini 3.1 Pro's tool prioritization and reduced hallucinations outpace prior models, narrowing the gap in enterprise dev tools. Amid rising demand for reliable AI assistants, it positions Google to capture more developer market share, especially versus Cursor's integrations, by enabling seamless, in-IDE large-scale edits without context limits.
