Agents & Models
Documentation and guides for OutcomeDev.
OutcomeDev enables you to use the world's most powerful AI models. Here is how to choose the right one for your needs.
Available Agents
Claude (Anthropic)
The Coding Specialist
- Best For: Writing clean code, refactoring, and understanding large projects.
- Models:
- Opus 4.6: Maximum intelligence for complex problems, with top-tier performance.
- Sonnet 4.6: Advanced for complex agents, excelling in computer use, coding, and reasoning. Also the default for Free/Pro.
- Haiku 4.5: Lightweight and fast, ideal for simple tasks and quick responses.
Codex (OpenAI)
The Reasoner
- Best For: Architecture, logic, and following complex instructions.
- Models:
- GPT-5.3 (Reasoning): The most advanced model for complex reasoning and problem-solving.
- GPT-5.3-Codex: Excels in code generation, understanding, and completion.
- GPT-5.3-Codex-Spark: Optimized for rapid code generation and quick iterations.
- GPT-5.3 Instant: Fast responses for quick tasks.
- GPT-5.2 (Thinking): "Thinks" before answering to solve hard logic puzzles.
- GPT-5.2 Instant: Fast responses for quick tasks.
Gemini (Google)
The Multimodal Expert
- Best For: Working with images, design, and massive amounts of data.
- Models:
- Gemini 3.1 Pro (Preview): Successor to Gemini 3 Pro with improved reasoning and token efficiency. Optimized for software engineering.
- Gemini 3 Pro (Preview): Our most advanced reasoning model for complex multimodal problems (1M context).
- Gemini 3 Flash (Preview): Agentic workhorse with near-Pro intelligence and balanced speed.
- Gemini 3.1 Flash-Lite (Preview): Best-in-class cost/performance for high-volume agentic tasks.
- Gemini 3 Pro Image (Preview): Specialized model for high-fidelity image generation and multimodal editing.
- Gemini 2.5 Pro: Reliable, low-latency reasoning for advanced production tasks.
- Gemini 2.0 Flash: Fast, multimodal model for general task completion.
How to Choose?
| Goal | Recommended Model |
|---|---|
| I need complex logic & architecture | Gemini 3.1 Pro (Preview) / GPT-5.3 (Reasoning) |
| I need high-quality code generation | Claude Opus 4.6 / Gemini 3.1 Pro (Preview) |
| I need complex agent capabilities/computer use | Claude Sonnet 4.6 / Gemini 3 Flash (Preview) |
| I need to work with designs/images | Gemini 3.1 Pro (Preview) / Gemini 2.5 Flash Image |
| I need a quick answer | Gemini 3.1 Flash-Lite (Preview) / GPT-5.3 Instant / Claude Haiku 4.5 |
Compare Mode
Not sure which model to use? Select the Compare agent. This allows you to run your prompt against multiple models simultaneously (e.g., Claude vs. GPT-5.2) so you can pick the best result.
Next Steps
Once you've chosen a model, you'll want to select the right framework for your project. Check out our Framework Guide to learn more.