Octen Outpaces Industry Giants in Autonomous AI Research Benchmarks

The DeepResearch Bench serves as the primary standard for evaluating how AI systems synthesize complex information across multiple sources. While competitors like OpenAI Deep Research and Gemini Deep Research often require significant processing time to generate findings, Octen’s architecture relies on massively parallel queries to complete tasks in two to three minutes. The platform achieved an overall benchmark score of 55.58, outperforming its closest rivals by margins ranging from 10 to 17 points.

Kuan "Colin" Zou, CEO and founder of Octen, attributes this performance to the system's foundational design, which prioritizes large-scale research speed without compromising depth. Beyond its recent performance in autonomous research, the company has maintained the top position on the Retrieval Embedding Benchmark since January. By providing a real-time, LLM-native search API, Octen aims to replace the complex, custom-built retrieval pipelines currently used by AI agents and research assistants.