{"slug": "perplexity-launches-a-local-ai-agent-with-zero-token-costs", "title": "Perplexity launches a local AI agent with zero token costs", "summary": "Perplexity launched 'Portable Computer,' a local-first AI agent developed with NVIDIA, which runs entirely on-device using small models such as Qwen 3.8 and NVIDIA Nemotron 3.5 Lightning, ensuring user data is not sent to the cloud. Tasks handled locally incur zero token costs, and the agent can connect to cloud models for web search and advanced reasoning with user permission. The tool is available starting today for Perplexity Pro or Max subscribers on NVIDIA DGX Spark systems.", "body_md": "Affiliate links on Android Authority may earn us a commission. [Learn more.](https://www.androidauthority.com/external-links/)\n\n# Perplexity launches a local AI agent with zero token costs\n\nAug 26, 2026 — 2:40 AM ET\n\n- Perplexity has launched a new agentic AI tool called “Portable Computer” that runs locally on users’ devices.\n- It uses small, efficient models such as Qwen 3.8 and NVIDIA Nemotron 3.5 Lightning.\n- It will not incur token costs for tasks completed locally, but can connect to cloud models if needed.\n\nMonths after launching its agentic AI tool, “Personal Computer,” [Perplexity](https://www.androidauthority.com/perplexity-usage-limits-promo-code-crackdown-3668441/) is introducing a new option for users who prefer to run their [AI agents locally](https://www.androidauthority.com/meta-muse-glimmer-3696698/).\n\nThe company today [announced](https://www.perplexity.ai/hub/blog/introducing-portable-computer-for-local-first-ai) “Portable Computer,” its local-first agent, in partnership with NVIDIA. Portable Computer will run entirely on-device. The agent, along with the models and other tools, lives locally on the user’s device. This ensures that user data isn’t sent to the cloud and offers a higher level of privacy for sensitive data.\n\nPortable Computer will use very small models such as Qwen 3.8 and NVIDIA Nemotron 3.5 Lightning. According to Perplexity, these models can perform complex agentic tasks and run on local hardware, making them a good choice for its new local agentic AI.\n\nHowever, Portable Computer still supports functions such as web search and advanced reasoning with cutting-edge models. If any such feature is required for the task at hand, Perplexity says the agent will always ask the user for permission before proceeding.\n\nThe company also states that work handled locally by the user’s machine will not incur any token costs. That could make a huge difference for users who need to carry out high-volume tasks without worrying about exhausting their tokens or incurring extra costs.\n\nPerplexity’s Portable Computer currently supports NVIDIA’s DGX Spark platform and can run Qwen 3.8 or PPLX 27B — a post-trained version of the Qwen 3.8 model. The company also says that support for the NVIDIA Nemotron 3.5 Lightning model, as well as NVIDIA RTX GPUs, is coming soon. Users subscribed to Perplexity Pro or Max can set up [Portable Computer](https://www.perplexity.ai/hub/products/portable-computer) on their [NVIDIA DGX Spark](https://www.nvidia.com/en-us/products/workstations/dgx-spark/) systems starting today.\n\nThank you for being part of our community. Read our [Comment Policy](https://www.androidauthority.com/android-authority-comment-policy/) before posting.", "url": "https://wpnews.pro/news/perplexity-launches-a-local-ai-agent-with-zero-token-costs", "canonical_source": "https://www.androidauthority.com/perplexity-portable-computer-local-ai-agent-3703083/", "published_at": "2026-08-26 06:40:55+00:00", "updated_at": "2026-08-26 06:45:14.946463+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-products", "ai-infrastructure"], "entities": ["Perplexity", "NVIDIA", "Qwen 3.8", "NVIDIA Nemotron 3.5 Lightning", "PPLX 27B", "NVIDIA DGX Spark", "NVIDIA RTX GPUs"], "alternates": {"html": "https://wpnews.pro/news/perplexity-launches-a-local-ai-agent-with-zero-token-costs", "markdown": "https://wpnews.pro/news/perplexity-launches-a-local-ai-agent-with-zero-token-costs.md", "text": "https://wpnews.pro/news/perplexity-launches-a-local-ai-agent-with-zero-token-costs.txt", "jsonld": "https://wpnews.pro/news/perplexity-launches-a-local-ai-agent-with-zero-token-costs.jsonld"}}