Peter Smith
Perplexity is stepping up its AI research game with a major backend upgrade. The platform's Deep Research feature has quietly switched to Anthropic's powerful Opus 4.6 model, delivering stronger performance for users who need comprehensive, multi-layered answers. Max subscribers already have access, while Pro users are seeing the update roll out gradually.
Perplexity Deep Research Gets Anthropic Opus 4.6 Under the Hood
As Perplexity reported, the company upgraded its Deep Research feature to Anthropic's Opus 4.6 model. The rollout started with Max-tier subscribers and is progressively expanding to Pro users.
According to Perplexity, the new model delivers measurable improvements across both internal testing and external benchmarks. The company shared evaluation results showing Perplexity Deep Research scoring 81.9% in performance metrics—outpacing several competing systems in the process.
The comparison chart positioned Perplexity ahead of Moonshot K2.5 at 77.1%, Anthropic Opus 4.5 at 76.1%, OpenAI GPT-5.2 at 71.3%, and Gemini Deep Research Agent at 66.1%. These numbers reflect the platform's ability to handle complex, multi-step research queries with greater accuracy.
"The change improves existing results across internal and external benchmarks," Perplexity stated in its announcement.
What This Means for Deep Research Users
The upgrade doesn't change subscription pricing or access tiers—it simply swaps out the engine powering research responses. Deep Research focuses on generating detailed, long-form answers by synthesizing information across multiple sources and reasoning steps.
For Pro and Max subscribers, this means more reliable outputs when tackling research-heavy questions that require connecting dots across various data points. The feature continues to operate within the same workflow, maintaining its focus on comprehensive information synthesis rather than quick, surface-level answers.
SEO: Anthropic Opus 4.6 vs Competing AI Research Models
The performance gap between Opus 4.6 and previous models like Anthropic Opus 4.5 (76.1%) shows incremental but meaningful progress in AI research capabilities. Meanwhile, the gap against OpenAI's GPT-5.2 (71.3%) and Google's Gemini Deep Research (66.1%) suggests Perplexity is betting on Anthropic's architecture for sustained advantage in the AI research assistant space.
As the rollout continues, Pro subscribers should expect access over the coming weeks, with the same subscription structure remaining in place throughout the transition.
Peter Smith
