#搜索

check out our new blog post on how we serve embeddings and rerankers for our SOTA pplx-embed models over an exabyte-scale search index. embedding models have a similar serving profile to LLMs: batch indexing is compute-bound prefill, online serving is memory-bound decode. this lets us reuse the same optimized kernels we built for large LLMs and get great efficiency for free. on top of that, we optimized the runtime: whole-model CUDA graphs captured lazily as the engine serves, and a LazyTensor in Rust that overlaps CPU scheduling with GPU execution. result: up to 3x lower p50 and 4.8x lower p99 latency than vLLM on BGE-M3 at 128 tokens, single H200! if you are interested in working on problems like this, DM me or apply at http://perplexity.ai/hub/careers

Every answer in Perplexity starts with embedding and ranking models picking the most relevant results for the query. Today we published research on how we built SoTA serving infrastructure behind those models. Read the research: https://www.perplexity.ai/hub/blog/fast-embeddings-on-gpus

我们在 WANDR 上评估了 GPT-6 Astra。它以每任务 $11.98 的成本得分 0.682，是我们测试过的所有模型中的最高分。 GPT-6-Astra 比 Fable 5.1 高 13.5%，成本低 6.1%；比 Opus 5 高 27.0%，成本高 3.3%。

Will it even work. 🤔 WSJ: New York City is pausing student-facing generative AI through eighth grade. The moratorium affects nearly 600,000 students, removing student-facing generative AI from those grades, and prohibits companion chatbots across all grades. High schools take a controlled route: up to 50,000 students can join 5 teacher-supervised AI pilots, while all high schoolers receive twice-yearly AI literacy instruction. to note, New York City public schools previously imposed a temporary ban on ChatGPT soon after its rollout in 2022. The restriction was later lifted and replaced by a custom AI-powered teaching assistant, And actually, there is also strong evidence on the other side. In a 2025 randomized Harvard physics study, 194 college students using a carefully designed AI tutor had more than twice the median learning gain of students receiving an active-learning classroom lesson, while typically spending less time.

DGX Spark Live: Perplexity Portable Computer Goes Local https://x.com/i/broadcasts/1jGXgBXzpOXKZ

Want to see @perplexity_ai Portable Computer running locally on NVIDIA DGX Spark? 👀 Tune in tomorrow, 9/2 at 12:30 p.m. PDT, to see live demos showcasing private file analysis, connected tools and more.

https://x.com/i/article/2094451432208711681

Perplexity Search debuts on the Artificial Analysis Search Index, with all three context size variants taking top positions on the leaderboard The @perplexity_ai Search API comes with three context settings (low, medium, and high) that control how much extracted content each search result carries. We tested all three variants using our standardized methodology: the same model (GPT-5.6 Luna at medium reasoning), running inside Stirrup, our open-source agent harness, with tools for searching and fetching pages from the web. Only the provider behind the search tool changes. Key results: ➤ Perplexity Search (medium) scores 80 on the Artificial Analysis Search Index, ahead of the previous leaders, Parallel (advanced) and Brave Search (LLM context), at 75. The high and low variants score 79 and 77 respectively. Its lead is concentrated in BrowseComp results, with AA-Omniscience and DeepSearchQA scoring comparably to other leading providers ➤ Efficient search payloads: smaller overall search results mean the model reads less per task, so Perplexity has the lowest model inference cost per task of providers we’ve tested so far, ranging from $0.028 to $0.034 across the three variants vs $0.036 for the next lowest provider ➤ Total cost per task is ~$0.091 for the medium and high context variants, at mid-pack latency. For comparison, Parallel (advanced) costs $0.084 per task and Brave (LLM context) costs $0.13 per task

Perplexity Search debuts on the Artificial Analysis Search Index, with all three context size variants taking top positions on the leaderboard The @perplexity_ai Search API comes with three context settings (low, medium, and high) that control how much extracted content each search result carries. We tested all three variants using our standardized methodology: the same model (GPT-5.6 Luna at medium reasoning), running inside Stirrup, our open-source agent harness, with tools for searching and fetching pages from the web. Only the provider behind the search tool changes. Key results: ➤ Perplexity Search (medium) scores 80 on the Artificial Analysis Search Index, ahead of the previous leaders, Parallel (advanced) and Brave Search (LLM context), at 75. The high and low variants score 79 and 77 respectively. Its lead is concentrated in BrowseComp results, with AA-Omniscience and DeepSearchQA scoring comparably to other leading providers ➤ Efficient search payloads: smaller overall search results mean the model reads less per task, so Perplexity has the lowest model inference cost per task of providers we’ve tested so far, ranging from $0.028 to $0.034 across the three variants vs $0.036 for the next lowest provider ➤ Total cost per task is ~$0.091 for the medium and high context variants, at mid-pack latency. For comparison, Parallel (advanced) costs $0.084 per task and Brave (LLM context) costs $0.13 per task

We’re partnering with @Perplexity_AI to bring live web search to Decagon agents. Customer questions often depend on information that changes by the hour. Agents can now search the live web mid conversation, pull in current information, and respond with cited sources.