What Perplexity Announced

On March 5, 2026 (UTC), Perplexity posted on X that GPT-5.4 and GPT-5.4 Thinking are now available to Pro and Max subscribers. The wording is direct: this is a live paid-tier availability update, not a preview waitlist message.

At crawl time using public mirror endpoints, the post had more than 1,500 likes and over 84,000 views, suggesting fast user attention from people who want frontier-model access inside a search-first product.

Why It Matters

Perplexity competes by combining web retrieval, source grounding, and LLM response generation in one workflow. Adding GPT-5.4 class models strengthens its premium position for users who compare model quality across multiple AI assistants.

For teams, the practical question is not just model branding. It is whether answer quality, reasoning depth, and time-to-output improve in daily research tasks. This update creates a clear test point for organizations already using Pro or Max seats.

What Is Still Unknown

The X post does not provide detailed benchmark methodology or per-task performance breakdowns. Teams considering broad adoption should run task-specific evaluations on accuracy, consistency, latency, and cost before standardizing workflows.

#llm-platform

Perplexity Adds GPT-5.4 and GPT-5.4 Thinking for Pro and Max Subscribers