Gemini 3.1 Pro Custom Tools is purpose-built for driving Atlas, the terminal-native AI coding agent, especially when interacting with its dense custom tool surface. Every Atlas action, from code edits to bash calls, is a tool invocation. This model offers a substantial 1,048,576 token context window and is priced at $2 per Mtok input and $12 per Mtok output, making it a powerful and cost-effective choice for developers in 2026.
Why choose Gemini 3.1 Pro Custom Tools for Atlas?
Gemini 3.1 Pro Custom Tools excels in Atlas because it is purpose-built for dense custom tool surfaces, where every Atlas action is a tool invocation. This model is a distinct registry entry, gemini-3.1-pro-preview-customtools, specifically designed for scenarios involving a high volume of tool calls.
Atlas, as a terminal-native AI coding agent, relies heavily on tool invocations for its core functionalities. Every edit, read, grep, and bash call executed by Atlas is a tool invocation. Gemini 3.1 Pro Custom Tools is a specialized variant of Gemini 3.1 Pro, shipped by Google with a tool-calling variant of the same checkpoint. This makes it uniquely suited to interpret and sequence the complex tool calls that Atlas generates, ensuring more precise and reliable agent behavior. It maintains the identical 1,048,576 token context window and 65,536 token output ceiling as the base 3.1 Pro checkpoint, meaning no capabilities are sacrificed for its tool path optimization. This focus on tool interaction is crucial for Atlas's ability to search code with Axis, the hybrid semantic and keyword code search fused by reciprocal rank fusion, index code by AST declarations using tree-sitter, and manage file changes as git patches.
What are the pricing and context window for Gemini 3.1 Pro Custom Tools?
Gemini 3.1 Pro Custom Tools provides a substantial 1M token (1,048,576) context window, identical to the standard Gemini 3.1 Pro, ensuring ample space for complex coding tasks. Its pricing is set at $2 per Mtok for input and $12 per Mtok for output, offering a consistent cost structure.
Developers in 2026 considering Gemini 3.1 Pro Custom Tools for Atlas will find its resource allocation generous and its pricing transparent. The model boasts a 1,048,576 token context window, which is a significant capacity for handling large codebases, extensive documentation, and detailed agent plans. This large context window is particularly beneficial for Atlas, which drafts a plan in a read-only plan agent and asks before switching to a build agent, and computes a unified diff for every file edit. The output ceiling is 65,536 tokens, providing ample room for detailed responses and code generation. Crucially, the pricing for Gemini 3.1 Pro Custom Tools is identical to the base 3.1 Pro: $2 per Mtok for input and $12 per Mtok for output. This means choosing the purpose-built tool variant incurs no additional cost compared to the standard model, allowing developers to optimize for tool performance without financial penalty.
What are the tradeoffs of Gemini 3.1 Pro Custom Tools?
While Gemini 3.1 Pro Custom Tools offers significant advantages for Atlas, developers should be aware of two primary tradeoffs in 2026. The first is the potential for confusion between its distinct registry ID, gemini-3.1-pro-preview-customtools, and the standard gemini-3.1-pro-preview.
The primary tradeoff when selecting Gemini 3.1 Pro Custom Tools is the existence of two near-identical IDs in the Google model registry. Developers must explicitly choose gemini-3.1-pro-preview-customtools over gemini-3.1-pro-preview. Picking the wrong ID will silently change tool behavior, potentially leading to suboptimal performance or unexpected agent actions within Atlas. The second tradeoff is its preview status. Google may, at any time, fold this specialized variant back into the base Gemini 3.1 Pro model. While this might simplify model selection in the future, it means developers should monitor announcements for potential changes to the model's availability or naming. Despite these considerations, the model's purpose-built nature for tool calling remains a strong advantage for Atlas users, who can mitigate these tradeoffs by carefully following setup instructions and reviewing Atlas's unified diffs before writing changes.
When should I use a different model instead of Gemini 3.1 Pro Custom Tools?
Developers might consider an alternative to Gemini 3.1 Pro Custom Tools if the model's preview status or the potential for ID confusion presents an unacceptable operational risk in 2026. If your Atlas workflow does not heavily rely on complex, dense tool surfaces, a different model might suffice.
While Gemini 3.1 Pro Custom Tools is optimized for Atlas's tool-heavy operations, there are scenarios where another model might be preferable. If the 'preview' status from Google is a concern for long-term stability or if the slight risk of confusing gemini-3.1-pro-preview-customtools with gemini-3.1-pro-preview is deemed too high, a more stable or less specialized model could be chosen. For Atlas users who primarily engage in simpler, less tool-intensive tasks, or those who prioritize a single, unambiguous model ID, the benefits of the custom tools variant might not outweigh these minor operational considerations. However, for any developer leveraging Atlas's full suite of capabilities, including its ability to read git branches, status, and diffs, and stage and create commits on your behalf, the tool-calling optimization of Gemini 3.1 Pro Custom Tools remains a significant advantage.
Setup
- 01Export your Google API key: Export GEMINI_API_KEY="YOUR_API_KEY_HERE"
- 02Alternatively, store the key securely: atlas login --provider google
- 03Verify model availability: Run atlas models google and look for gemini-3.1-pro-preview-customtools, distinct from gemini-3.1-pro-preview.
- 04Pin the model in your Atlas configuration: Add "model": "google/gemini-3.1-pro-preview-customtools" to your atlas.json file.
- 05Start with Atlas permissions in ask mode: Configure permissions to ask (allow/ask/deny) initially to observe how the tool variant sequences calls.
- 06Review proposed changes: Always review the unified diff Atlas shows before each write operation to sanity check the tool path and ensure desired outcomes.
Frequently asked questions
- What is the main advantage of Gemini 3.1 Pro Custom Tools for Atlas?
- The main advantage is its purpose-built design for custom tool surfaces. Every Atlas action, such as an edit or a bash call, is a tool invocation, and this model is optimized to handle these dense tool calls more effectively than standard models.
- How large is the context window for Gemini 3.1 Pro Custom Tools?
- Gemini 3.1 Pro Custom Tools features a substantial 1,048,576 token context window, which is equivalent to 1M tokens. This large capacity allows Atlas to process extensive codebases and complex instructions.
- What is the pricing for Gemini 3.1 Pro Custom Tools?
- The pricing for Gemini 3.1 Pro Custom Tools is $2 per Mtok for input tokens and $12 per Mtok for output tokens. This cost structure is identical to the base Gemini 3.1 Pro model.
- Are there any differences in capabilities between this model and standard Gemini 3.1 Pro?
- No, there are no capabilities traded away. Gemini 3.1 Pro Custom Tools shares the identical 1,048,576 token context and 65,536 token output ceiling with the base 3.1 Pro checkpoint, but it is specifically optimized for tool-calling.
- What is the primary tradeoff when using Gemini 3.1 Pro Custom Tools?
- The primary tradeoff is the potential for confusion between its distinct registry ID, gemini-3.1-pro-preview-customtools, and the standard gemini-3.1-pro-preview. Selecting the wrong ID will silently alter tool behavior.
- Is Gemini 3.1 Pro Custom Tools a stable release?
- Gemini 3.1 Pro Custom Tools is currently in preview status. This means Google may, at any time, fold this specialized variant back into the base Gemini 3.1 Pro model, so developers should stay informed about its status.
- How do I ensure Atlas uses the correct Gemini 3.1 Pro Custom Tools model?
- You must explicitly pin "model": "google/gemini-3.1-pro-preview-customtools" in your atlas.json configuration. It is also recommended to verify its presence using atlas models google.
Try SeaShell in your terminal
The terminal-native AI coding agent. Free core, single binary.
Install SeaShellRelated guides
Atlas for Bash in 2026
Atlas is a terminal-native AI coding agent for Bash in 2026, with shellcheck-aware suggestions. Harden a script or fix shellcheck warnings and review every diff.
Atlas vs Mistral Vibe for Code: Terminal AI Coding Agents in 2026
Compare Atlas and Mistral Vibe for Code in 2026. Atlas offers terminal-native TUI, permission-gated tools, and local embeddings. Mistral Vibe provides a four-model stack and EU data sovereignty.
Atlas for Bun: Terminal-Native AI Coding in 2026
Atlas is a terminal-native AI coding agent for Bun. Swap Node compatibility shims for Bun.serve and bun:sqlite, iterate with bun test, and format with bun fmt in 2026.
Atlas for Express in 2026
Atlas is a terminal-native AI coding agent for Express in 2026. It reads your middleware order, wraps async route handlers, and runs mocha behind a prompt.
Atlas vs OpenAI Codex CLI: A Developer's Guide for 2026
Compare Atlas and OpenAI Codex CLI for 2026. Atlas offers terminal-native TUI, BYO models, and granular change review, while OpenAI Codex CLI provides sandboxed execution and OpenAI model integration.
Atlas for Next.js in 2026
Adopt Atlas, the terminal-native AI coding agent, for Next.js development in 2026. Enhance productivity across App Router, server components, and API routes with secure, reviewable AI assistance.
Atlas for Groovy: A Terminal-Native AI Coding Agent for Gradle, Spock, and Jenkins in 2026
Atlas is a terminal-native AI coding agent for Groovy in 2026. It reads build.gradle closures and Jenkinsfiles, writes Spock specs, runs ./gradlew test, and applies Spotless.
Atlas for Actix Web in 2026
Atlas is a terminal-native AI coding agent for Actix Web in 2026. It reads extractors and app_data, then runs cargo test and cargo clippy behind a prompt.