Perplexity says its API stack now spans agent orchestration, real-time search, embeddings, and an upcoming sandbox under one platform. The update packages more of the agent runtime into Perplexity infrastructure instead of leaving developers to assemble separate providers.
Perplexity’s Computer account used X on March 9, 2026 to demonstrate Claude Code and GitHub CLI running directly inside Perplexity Computer. In the public demo, the system forked an Openclaw repository, planned a fix, implemented the change, and submitted a pull request from inside the Computer environment.
Perplexity says users can now guide Perplexity Computer by voice, not just text. The update turns mid-task feedback and redirection into a spoken control loop for long-running agent work on the web.
Perplexity announced on March 5, 2026 that GPT-5.4 and GPT-5.4 Thinking are now available for Pro and Max subscribers. The move strengthens paid-tier access to frontier LLM options.
Perplexity announced Ask 2026 on X on February 27, 2026 (UTC), describing it as the company’s inaugural developer conference. The event page specifies an invitation-only format in San Francisco on March 11 with a limited builder-focused agenda.
Perplexity announced on February 26, 2026 that Samsung will ship Galaxy S26 devices with Perplexity integrated at the system level. The company also says the assistant will support a dedicated wake phrase, “Hey Plex.”
Perplexity announced on February 26, 2026 that `pplx-embed-v1` and `pplx-embed-context-v1` are now available in 0.6B and 4B variants. The company positions the release as retrieval-first infrastructure with quantized embeddings and benchmark-focused performance claims.
Perplexity said on 2026-02-26 that Galaxy S26 devices will ship with Perplexity integrated as a system-level AI. In follow-up posts, the company also described API integration with Bixby and additional Samsung Internet and default-search distribution plans.
Perplexity announced Perplexity Computer on 2026-02-25 and described it as a system that can research, design, code, deploy, and manage projects end-to-end. In the same X thread, the company said the product routes work across 19 models and launches first for Max subscribers on web.
Perplexity launched Model Council, a system running multiple frontier AI models including Claude, GPT-5.2, and Gemini in parallel to generate unified, cross-validated answers. This significantly improves reasoning quality and reduces hallucination errors.