The Clarity platform offers multiple agents, enabling you to experiment with agents that use different AI models from different AI companies and choose the appropriate model for your use case. This page provides an overview of the agents available on the Clarity platform to help you understand their capabilities, differences, and how to choose the right one for your needs.
OpenAI Models
GPT-4o agent
The default agent in Clarity uses OpenAI’s GPT-4o model, an advanced large language model developed to handle complex, multi-step tasks with a high degree of accuracy. GPT-4o excels at generating nuanced and contextually relevant responses, making this agent suitable for sophisticated applications and datasets.
GPT-4o mini agent
This agent uses Open AI’s GPT-4o mini model, a more compact and cost-effective version of GPT-4o. It offers a balance between performance and resource efficiency, making it a significantly more environmentally friendly choice while maintaining robust performance. GPT-4o mini is optimal for basic and lightweight tasks and surpasses other small models on academic benchmarks, while supporting the same range of languages as GPT-4o. For a more detailed breakdown of the model evaluation benchmark and additional information on GPT-4o mini, please refer to OpenAI’s documentation.
o3 agent
o3 is the most powerful reasoning model currently available from OpenAI and is intended for tasks requiring deep analytical thinking, problem-solving and complex reasoning. It takes longer to respond to queries and users more energy due to its longer responses. This agent will support file uploads in a future release.
| Feature | GPT-4o | GPT-4o mini | o3 |
|---|---|---|---|
| Performance[i] | Higher accuracy, coherence, and fluency | Lower accuracy, coherence, and fluency | Higher accuracy, coherence, and fluency |
| Environmental Impact | Higher energy consumption due to greater computational requirements | Lower energy consumption, making it more eco-friendly | Higher energy consumption due to greater computational requirements |
| Use Cases | Suitable for complex queries, creative writing, and detailed analysis | Geared towards basic tasks, and lower-resource applications | Suitable for complex queries, reasoning and complex analysis |
| Image Generation | Yes, using DALL·E 3 | Yes, using DALL·E 3 | No |
| Code execution | Can execute code to create charts or calculate answers | Can execute code to create charts or calculate answers | No (cannot execute code to provide analysis but can output code when prompted) |
| File uploads | Yes | Yes | No |
| Knowledge cutoff | October 2023 | October 2023 | May 2024 |
[i]LLM Stats provides scores for these models used by these agents on common AI benchmarks.
Anthropic Models
Claude 3.5 Sonnet agent
This agent uses Anthropic’s Claude 3.5 Sonnet model to generate text responses and generate code and can generate images using DALL·E 3. The Claude 3.5 Sonnet model is suitable for coding tasks, summarizing and reasoning on large amounts of text, and supporting creative and academic writing. The Claude 3.5 Sonnet agent will support file uploads in future releases.
Claude Sonnet 4 agent
This agent is very similar to the Claude 3.5 Sonnet agent but uses the latest Claude Sonnet model and does not generate images. The Claude Sonnet 4 model is suitable for coding tasks, summarizing and reasoning on large amounts of text, and supporting creative and academic writing. The Claude Sonnet 4 agent will support file uploads in future releases.
| Feature | Claude 3.5 Sonnet | Claude Sonnet 4 |
|---|---|---|
| Performance[ii] | High accuracy, coherence, and fluency | Higher accuracy, coherence, and fluency |
| Environmental Impact | Higher energy consumption due to greater computational requirements | Higher energy consumption due to greater computational requirements |
| Use Cases | Suitable for complex queries, creative writing and code generation | Suitable for complex queries, creative writing and code generation |
| Image Generation | Yes, using DALL·E | No |
| Code execution | No (cannot execute code to provide analysis but can output code when prompted) | No (cannot execute code to provide analysis but can output code when prompted) |
| File uploads | No | No |
| Knowledge cutoff | April 2024 | January 2025 |
[ii]LLM Stats provides scores for these models used by these agents on common AI benchmarks.
Google Models
Gemini 2.5 Pro agent
The Gemini 2.5 Pro model is designed for advanced reasoning and complex problem solving. With a 1-million token context window it can handle large amounts of data and documents. Gemini 2.5 Pro does sacrifice some speed in responding for more thinking. The Gemini 2.5 Pro agent will support file uploads in future releases.
Gemini 2.5 Flash agent
Gemini 2.5 Flash also has 1-milliion token context window, but is a lower latency model, so it responds more quickly and is suitable for chatting and quick responses, uses less energy, and is less expensive than Gemini 2.5 Pro. This agent will support file uploads in future releases.
| Feature | Gemini 2.5 Pro | Gemini 2.5 Flash |
|---|---|---|
| Performance[iii] | Higher accuracy, coherence, and fluency | Lower accuracy, coherence, and fluency |
| Environmental Impact | Higher energy consumption due to greater computational requirements | Lower energy consumption, making it more eco-friendly |
| Use Cases | Suitable for complex queries, creative writing, and detailed analysis | Geared towards basic tasks, and lower-resource applications |
| Image Generation | No | No |
| Code execution | No (cannot execute code to provide analysis but can output code when prompted) | No (cannot execute code to provide analysis but can output code when prompted) |
| File uploads | No | No |
| Knowledge cutoff | January 2025 | January 2025 |
[iii]LLM Stats provides scores for these models used by these agents on common AI benchmarks. Note that this comparison is to a preview model of Genesis 2.5 Pro, not the version available in Clarity.