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 featured agents available on the Clarity platform to help you understand their capabilities, differences, and how to choose the right one for your needs.

The 🌱 emoji indicates an agent which uses a more environmentally friendly model.

OpenAI Models

GPT-5.2 agent

The default agent in Clarity uses OpenAI’s GPT-5.2 model, an advanced large language model developed to handle complex, multi-step tasks with a high degree of accuracy. GPT-5.2 excels at generating nuanced and contextually relevant responses, making this agent suitable for sophisticated applications and datasets.

GPT-5 mini agent 🌱

This agent uses Open AI’s GPT-5 mini model, a more compact and cost-effective version of GPT-5. It offers a balance between performance and resource efficiency, making it a significantly more environmentally friendly choice while maintaining robust performance. GPT-5 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-5.2. 

o3 agent

o3 is a powerful reasoning model 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.

Feature GPT-5.2 GPT-5 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 No No No
Code execution Can execute code to create charts or calculate answers Can execute code to create charts or calculate answers Can execute code to create charts or calculate answers
File uploads  Yes Yes Yes
Knowledge cutoff August 2025 May 2024 May 2024

[i]LLM Stats provides scores for these models used by these agents on common AI benchmarks.

Anthropic Models

Claude Sonnet 4.6 agent

This agent uses Anthropic’s Claude Sonnet 4.6 model to generate text responses and code. The Claude Sonnet 4.6 model is suitable for coding tasks, summarizing and reasoning on large amounts of text, and supporting creative and academic writing.

Feature Claude Sonnet 4.6
Performance[ii] Higher accuracy, coherence, and fluency
Environmental Impact Higher energy consumption due to greater computational requirements
Use Cases Suitable for complex queries, creative writing and code generation
Image Generation No
Code execution Yes
File uploads  Yes
Knowledge cutoff August 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.

Gemini 2.5 Flash agent 🌱

Gemini 2.5 Flash has a 1 million-token context window and lower latency, enabling quicker responses and suitability for chatting. Additionally, it consumes less energy and is more cost-effective than Gemini 2.5 Pro.

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 Can execute code to create charts or calculate answers Can execute code to create charts or calculate answers
File uploads  Yes Yes
Knowledge cutoff January 2025 January 2025

[iii]LLM Stats provides scores for these models used by these agents on common AI benchmarks