AI SuperAgent
The World’s First Compound AI System
SuperAgent Suite
Intelligent AI models with low latency and affordable prices through our Compound AI System.
Unlike conventional AI models, each SuperAgent leverages a mixture of advanced AI models and cutting-edge research to deliver superior performance across diverse tasks, such as Coding and Deep Research.
The SuperAgent Suite
SuperAgent was designed to surpass the limitations of available models on the market. We often observed users comparing different models or engaging in multiple back-and-forth conversations to find the best answer. We wanted to solve these challenges to deliver the most comprehensive and accurate answers tailored to your needs—right from the first interaction, with unmatched speed and affordability.
For unmatched depth and precision.
Apex combines multiple flagship AI models for expert-level insights and is available to Ultra and Business subscribers.
For lightning-fast responses.
Turbo delivers instant responses using our custom in-house fine-tuned models and is available to all subscribers.
SuperAgent was designed to surpass the limitations of available models on the market. We often observed users comparing different models or engaging in multiple back-and-forth conversations to find the best answer. We wanted to solve these challenges to deliver the most comprehensive and accurate answers tailored to your needs—right from the first interaction, with unmatched speed and affordability.
For complex problems that require advanced reasoning.
SuperAgent-R is built on DeepSeek R1 distilled on Llama 70B, and is available to Ultra and Business subscribers.
SuperAgent-R 2 was built to advance reasoning capabilities across math, science, coding, and more. Current models have many limitations. They’re costly and unsustainable to run for businesses and customers. Or lack the necessary customization, features, and performance to power agentic workflows. Hence, we developed a reasoning model that’s fine-tunable, fast, and affordable for everyone.
For the most complex research and expert level insights.
Our AI research assistant crafts and executes plans that evolve as it learns new information.
At Ninja, we set out to give everyone their own personal AI assistant. One that goes beyond simple requests and can autonomously interact and complete tasks on your behalf. That's why we've innovated agentic workflows powered by reasoning models and tools. We started this journey with our Deep Research feature, and soon will add many more.
SuperAgent Benefits
Performance on-par with leading models. SuperAgent ensures you receive well-researched, well-rounded answers tailored to your prompt.
By drawing from the unique strengths of multiple models, SuperAgent provides answers that are more nuanced and comprehensive than what a single model could achieve.
Unlike other products on the market, our Ultra and Business users get unlimited access to SuperAgent capabilities—starting at just $15/month
Our SuperAgent Models provide immediate and more precise results compared to other products on the market. Now, you never have to wait long to get the answers you need.
Now available is Deep Research, an AI assistant designed to think critically and evolve its research strategy as it gathers information. Unlike simple AI assistants that return surface-level results, Deep Research is built to analyze, adapt, and refine its approach to deliver high-quality expert-level insights.
SuperAgent Use Cases
Generate optimized code snippets, detect bugs and enhance code quality.
Perform financial analysis by combining data from multiple sources, such as earnings reports, economic indicators, and sector trends. Identify hidden patterns and provide expert-level analysis.
Develop campaign ideas, strategies, and tactics informed by insights from consumer behavior, industry trends, and competitor analysis. Or simplify complex concepts, making them more accessible and compelling.
Craft clear, professional responses to customer inquiries and support tickets that match customer sentiment, address concerns, and enhance overall satisfaction.
How SuperAgent Compares to Other Models
SuperAgent Turbo & Apex Flagship Model
SuperAgent Apex scored the highest on the industry-standard Arena-Hard-Auto (Chat) test. It measures how well AI can handle complex, real-world conversations, focusing on its ability to navigate scenarios that require nuanced understanding and contextual awareness.
The models also excel in other benchmarks: Math-500, AIME2024 - Reasoning, GPQA - Reasoning, LiveCodeBench - Coding, and LiveCodeBench - Coding - Hard.
Last updated: 04/15/2025
Last updated: 04/15/2025
Last updated: 04/15/2025
Last updated: 04/15/2025
Last updated: 04/15/2025
Last updated: 04/15/2025
SuperAgent-R 2.0 Reasoning Model
SuperAgent-R 2.0 outperformed OpenAI O1 and Sonnet 3.7 in competitive math on the AIME test. It assesses AI’s ability to handle problems requiring logic and advanced reasoning.
SuperAgent-R 2.0 also surpassed human PhD-level accuracy on the GPQA test. It evaluates general reasoning through complex, multi-step questions requiring factual recall, inference, and problem-solving.
.avif)
Last updated: 04/15/2025
.avif)
Last updated: 04/15/2025
.avif)
Last updated: 04/15/2025
SuperAgent Deep Research
Deep Research achieved 91.2% accuracy on the SimpleQA test. It’s one of the best proxies for detecting the hallucination levels of a model. This highlights Deep Research’s exceptional ability to accurately identify factual information—surpassing leading models in the field.
In the GAIA test, Deep Research scored 57.64%, which indicates superior performance in navigating real-world information environments, synthesizing data from multiple sources, and producing factual, concise answers.
Deep Research also achieved a significant breakthrough in AI with a 17.47% score on the HLE test. It’s widely recognized as a rigorous benchmark for evaluating AI systems across more than 100 subjects. Deep Research performed notably higher than several other leading AI models, including o3-mini, o1, and DeepSeek-R1.
Last updated: 04/15/2025
Last updated: 04/15/2025
Provider (Pass @1)
Level 1
Level 2
Level 3
Average
OpenAI's Deep Research
74.29
69.06
47.6
67.36
Ninjas's Deep Research
69.81
56.97
46.15
57.64
Data source: OpenAI Blog post – Read more
%20Benchmark.avif)
Last updated: 04/15/2025
Frequently Asked Questions
Here's what you need to know about Ninja's SuperAgent based on what we get asked the most.
What is Compound AI?
Compound AI is a technology that leverages a mixture of advanced AI models to deliver superior performance across diverse tasks, such as coding and research. It sends your prompt to the most powerful AI models, critiques their responses, and then delivers more comprehensive, accurate, and helpful answers tailored to your needs.
Why is using multiple models better than just one?
Each AI model has unique strengths and specialties. By combining responses from multiple models, the SuperAgent provides richer perspectives, enhanced problem-solving capabilities, and more nuanced answers—all within a single interface.
Who can access the SuperAgent?
All paid users have access to SuperAgent Turbo. However, only Ultra and Business subscribers have access to Apex, SuperAgent-R 2.0, and Deep Research.
Can I customize which models the SuperAgent uses?
No, the choice of models is determined by Ninja based on a thorough analysis of each model’s capabilities. This ensures that you receive the most accurate and relevant answers.
