Perplexity’s Deep Research Surpasses OpenAI and Anthropic in Benchmark Tests

Perplexity has announced an advanced version of Deep Research that surpasses both OpenAI and Anthropic in benchmarking tests. This achievement highlights Perplexity’s focus on incorporating enhanced multi-step reasoning within Deep Research to deliver more comprehensive responses. The release comes amidst a broader industry trend where AI companies are rapidly developing specialized tools designed to outperform general-purpose models in specific tasks.

Source

You might be interested in …

Bitcoin Desk - The Bitcoin Street Journal cyberpunk, trending on artstation in the style of cyberpunk

AI & Tech Desk February 19, 2026

Grok releases 4.20 with 16 specialized AI agents

Grok has released its latest public test version, Grok 4.20, which utilizes 16 specialized AI agents designed to collaborate on generating responses in various domains, including coding, finance, physics, and ethics. This release reflects xAI’s […]

AI & Tech Desk March 1, 2026

AI systems face growing risks of silent failure, experts warn

The rise of artificial intelligence (AI) in business operations is unveiling significant risks, particularly the phenomenon of “silent failures at scale,” which could disrupt the economy. As AI systems become increasingly complex, organizations often struggle […]

AI & Tech Desk February 10, 2026

Kilo launches CLI 1.0 with transparent pricing to rival opaque subscription models

Kilo, a remote-first AI coding startup backed by GitLab co-founder Sid Sijbrandij, has launched Kilo CLI 1.0, a revamped command-line tool supporting over 500 AI models, including those from Alibaba. This release, positioned against IDE-centric […]