C++ scope
Limit each task to a named module, class, function, error, or test. Require include, lifetime, threading, reflection, ownership, and version assumptions to be stated.
Unreal programming workflow · C++ and Blueprint review
Gemini 3.6 Flash for Unreal C++ and Blueprints
The strongest safe use is not “write my whole game.” It is a narrow engineering loop: define ownership, supply the relevant C++ or Blueprint evidence, request one reviewable change, and prove the result with native Unreal tools.
Direct answer
For C++, ask Gemini 3.6 Flash for a minimal change plan and reviewable diff tied to exact engine and module context. For Blueprints, provide exported text, screenshots, variable and event descriptions, and the expected state transition. Never accept either result until compilation, runtime behavior, automation, replication, save/load, and packaged builds have been checked.
What this means for an Unreal team
Blueprint scope
Describe the graph as observable state: inputs, events, variables, authority, transitions, outputs, and failure behavior. Screenshots alone can hide pins, defaults, macros, and parent-class behavior.
Review contract
The model proposes; source control, compiler output, Unreal Header Tool, editor diagnostics, automation, packaged tests, and human reviewers decide.
Recovery contract
Every accepted task needs a clean baseline, small commit, failure capture, revert path, and repeated test after editor restart or packaged execution.
C++ prompt contract that reduces speculative edits
Provide the engine version, target platform, module Build.cs boundary, affected header and source files, relevant reflection macros, compiler or Unreal Header Tool output, and the smallest expected behavior. Ask the model to separate facts read from the supplied code from assumptions about engine APIs. Require it to explain ownership, lifetime, thread, replication, and garbage-collection risks before proposing code.
A good response should name the minimal files to change, show why each edit is necessary, list compile and runtime checks, and state what evidence would invalidate the plan. Avoid repository-wide rewrites, invented engine symbols, version-agnostic includes, or changes that suppress warnings without explaining the underlying state.
Blueprint evidence that a text model can review
Use Blueprint screenshots as orientation, not as the only source of truth. Include copied node text where available, parent class, interfaces, component hierarchy, variable types and defaults, event order, network authority, latent actions, timers, save behavior, and the exact failing path. For a large graph, divide it by responsibility and identify the entry and exit state for each slice.
Ask Gemini to return a graph plan rather than pretend it compiled a Blueprint. The plan should list nodes or functions, pin-level data flow, state invariants, invalid inputs, server-client ownership, and tests. A developer then implements the graph, compiles it, checks warnings, runs the behavior, restarts the editor, and validates a packaged build.
Regression sequence before accepting the change
- Compile the affected C++ targets and Blueprint assets from a clean baseline.
- Exercise the normal input, an invalid input, repeated input, interruption, level reload, and application restart.
- For multiplayer, test server authority, owning and simulated clients, reconnect, late join, and correction behavior.
- For persistence, test new saves, old saves, missing fields, corrupted data, and version migration.
- Capture the first failure, revert the model-authored change, and confirm the baseline returns before iterating.
Official evidence and capability boundary
Google's July 21, 2026 release is the source for model positioning, reported benchmarks, listed pricing, and availability. Google does not claim native Unreal integration on that page. Epic documentation and the target project remain authoritative for engine behavior.
Google release
Release date, positioning, reported efficiency, benchmark comparisons, pricing, and starting availability.
Open official announcementGemini model documentation
Re-check the exact model ID, supported inputs, current status, limits, API behavior, region, and terms before use.
Open model documentationGoogle DeepMind model card
Review evaluation scope, safety information, known limitations, and the evidence behind general capability claims.
Open model cardEpic documentation
Validate C++, Blueprint, testing, packaging, and runtime behavior in the exact Unreal version used by the project.
C++ documentation · Blueprint documentation · Automation tests
Continue through the Gemini 3.6 Unreal cluster
Gemini 3.6 Flash × Unreal
Evaluate Gemini 3.6 Flash for Unreal Engine planning, C++, Blueprints, multimodal review, cost, testing, and safe handoff without claiming native UE integration.
Read this guideGemini 3.6 vs 3.5 Flash
Compare Gemini 3.6 Flash and 3.5 Flash for Unreal coding, multimodal review, token efficiency, cost, migration, and controlled project evaluation.
Read this guideGemini 3.6 Multimodal Debugging
Build a safe Gemini 3.6 Flash workflow for Unreal screenshots, logs, traces, Blueprint evidence, rendering defects, and reproducible native validation.
Read this guideFAQ
Can Gemini 3.6 Flash compile Unreal C++?
The model can draft or review source text, but a response is not compiler evidence. Compile with the actual Unreal toolchain and target configuration, inspect Unreal Header Tool and linker output, run automation, and test the packaged target. Keep generated changes in source control so reviewers can attribute and revert them.
Can it read an Unreal Blueprint?
It can analyze screenshots, copied node text, exported descriptions, logs, and state diagrams supplied to it. That is not the same as loading the asset, resolving hidden defaults, compiling the graph, or executing it. Provide explicit graph context and validate every proposal inside the target Unreal project.
Should I convert Blueprint code to C++ automatically?
Not by default. Blueprint and C++ ownership should follow iteration needs, performance evidence, team skills, networking, maintenance, and API stability. Use the model to compare a specific system under measured constraints, then migrate only the responsibility supported by profiling, tests, and a reversible implementation plan.
Where does SEELE AI fit?
Use SEELE AI before implementation to make the player-facing interaction, camera, scene, and completion flow concrete in a browser-playable prototype. Carry the approved behavior into the native engineering backlog, but do not treat the browser prototype as compiled Blueprint, C++, or a packaged Unreal build.
Turn the research into a playable direction
Return to the Unreal landing page, choose a verified Workspace card, and make the scene or gameplay loop concrete before planning native implementation.