← Monday, November 16 deck

09Research

Perplexity Launches Frontier Research to Build an Automated Researcher on 10,000 GPUs

Today we're launching Perplexity Frontier Research. The goal of Perplexity Frontier Research is to create a system that can perform autonomous research. The work will focus on multi-agent orchestration, reinforcement learning, continual learning, systems, and pre-training. We will develop algorithms, models, and evaluations, released openly, and tested in Perplexity products. Our work will be supported by more than 10,000 GPUs. The Frontier Research team will regularly publish papers, models, data, and benchmarks. We're actively hiring researchers and engineers for Perplexity Frontier Research in San Francisco, Palo Alto, New York, and London.

A video frame from Perplexity's announcement of Frontier Research.

Context

Why it matters

A well funded search company is now spending a large GPU budget on open research toward an automated researcher, which adds another lab to the race to make agents that can run scientific work on their own.

The official Perplexity account posted the launch on November 16 and said the goal is a system that can perform autonomous research, supported by more than 10,000 GPUs. Chief executive Aravind Srinivas quoted that post and added that the company will publish evaluations, papers and blogs and will open source models and tooling. The official post is used as the lead because it is the company account and it cleared the 100,000 view floor.

Deep dive

Perplexity said it is launching Perplexity Frontier Research with the goal of creating a system that can perform autonomous research. The official post said the work will focus on multi agent orchestration, reinforcement learning, continual learning, systems, and pre training. The company said it will develop algorithms, models, and evaluations, release them openly, and test them in Perplexity products, supported by more than 10,000 GPUs. It also said the team will regularly publish papers, models, data, and benchmarks, and that it is hiring researchers and engineers in San Francisco, Palo Alto, New York, and London.

Srinivas's quote post restated the GPU commitment as at least 10,000 GPUs and said the company will publish evaluations, papers and blogs and open source its models and tooling. Those are company promises, not results already in hand. The launch site is research.perplexity.ai.

This is a research lab announcement, not a released automated researcher. No model weights, evaluation scores, or papers were published in the launch posts themselves. The 10,000 GPU figure is a resource commitment from Perplexity, not an independent count of current cluster size.

What to expect next

The first useful signals will be the papers, models, data, and evaluations Perplexity said it will release, and whether those artifacts are in fact open. Watch also for hiring updates and for any sign of the work appearing inside Perplexity's consumer products. Until a system is demonstrated on public benchmarks, treat automated researcher as a goal rather than as a present capability.