Available Models
PhoenixAI offers three models, each suited to different types of tasks. GPT-5.4 is the default model for general use, Opus-4.7 is the premium model for complex reasoning and coding, and Sonnet-4.6 provides a balance of speed and intelligence for everyday work.
GPT-5.4 (OpenAI) Features
- The standard model automatically selected when opening the PhoenixAI tool.
- A prompt limit of 75 allows users to ask frequent questions and gather information.
- Prompts are evaluated once and the resulting answers are provided to the user.
- Users may upload files to ask questions about.
- Users may create custom Assistants for specific localized sources of information.
Recommendation: This model is most useful to anyone who has a large variety of questions that need to be answered quickly. This model is great for most tasks and general chatting.
Opus-4.7 (Claude) Features
- Alternative model that must be selected via the dropdown menu.
- A prompt limit of 20.
- Prompts are evaluated with adaptive reasoning, allowing the model to think as much or as little as a task requires before responding.
- Users may upload files to ask questions about, including high-resolution images, which it reads in sharp detail.
- Users may create custom Assistants for specific localized sources of information.
- Handles complex, long-running tasks with rigor and consistency, paying precise attention to instructions and verifying its own outputs before responding.
- Optimized for advanced software engineering, math, and complex multi-step reasoning.
Recommendation: This model is most useful to anyone who has limited, high-stakes questions that require precise, well-reasoned answers. This model prioritizes accuracy and depth on the hardest coding and reasoning work.
Sonnet-4.6 (Claude) Features
- Alternative model that must be selected via the dropdown menu.
- A prompt limit of 75, aligning with GPT-5.4.
- Prompts are evaluated quickly, delivering faster responses than Opus while retaining strong reasoning.
- Users may upload files to ask questions about, with strong performance reading enterprise documents such as charts, PDFs, and tables.
- Users may create custom Assistants for specific localized sources of information.
- Offers near-Opus capability at greater speed and lower cost, making it well suited to everyday work at scale.
- Optimized for coding, document analysis, and general-purpose agentic tasks.
Recommendation: This model is most useful to anyone who wants a strong balance of intelligence, speed, and cost for a high volume of everyday tasks. This model is a dependable daily driver that handles most work without reaching for the premium model.
Change the Default Model
GPT-5.4 is automatically selected as the default model when opening PhoenixAI. To switch to Opus-4.7 or Sonnet-4.6, click the dropdown menu at the top of the screen labeled with the name of the current model and select your preferred option. The selected model will remain active for that conversation. Starting a new chat will reset the selection back to GPT-5.4.