Perplexity releases Photon engine and affordable Fast Search

Perplexity introduced the Rust-based Photon search engine and a Fast Search mode in the Search API for $1 per 1,000 requests, where, according to the company, 95% of results are returned in 230 ms or faster with slightly lower result quality.

Compiled by Sergey KostenchukPublished 2026-10-18Updated 2026-10-19

2026-10-18 Coding Perplexity

💻 Perplexity releases Photon engine and affordable Fast Search

The Rust-based search engine finds and ranks web pages before the language model assembles the answer. Based on it, a Fast Search mode appeared in the Search API — $1 per 1,000 requests instead of $5 for standard search.

🌍 According to Perplexity, 95% of results are returned in 230 ms or faster (median 160 ms), and on six agentic benchmarks, the fast mode scored 64.3% compared to 64.0% for the standard mode. For agentic pipelines, this is search five times cheaper with nearly the same quality.

👤 The mode is enabled with the parameter search_type: "fast" (in the Python SDK — via extra_body), with max_results from 1 to 20. The trade-off is honest: result quality is slightly lower (DCG dropped from 2.45 to 2.21), so leave standard web search for complex questions.

Source 1: https://community.perplexity.ai/t/introducing-fast-search-in-the-perplexity-search-api/6195 Source 2: https://docs.perplexity.ai/docs/search/fast-search

Sources