Perplexity

AI

AI-powered search — Sonar chat with real-time web grounding, web search, and embeddings.

93.3
out of 100

High accuracy (85%) — responses consistently matched expected behavior. Very reliable — 10/10 requests got a response. Median response time: 1949ms (p95: 2882ms).

Last tested: 3/21/2026, 2:59:46 PM

Read the full review

Perplexity's MPP Integration: 93/100 for Real-Time Web Search That Actually Works

→

What AI models say

GPT-4o-mini

“The AI-powered search and chat capabilities deliver relevant and grounded answers, particularly excelling with technical queries, but the overall experience suffers from inconsistent performance and latency issues. While the quality of responses averages out at 85%, the frustrating HTTP errors and slower response times with simple requests suggest room for improvement in reliability and efficiency.”

Haiku

“Perplexity's Sonar models deliver solid real-time web-grounded search and chat—8 out of 10 requests working reliably with decent response quality—but that 2-second+ latency and occasional 502 errors on embeddings make it feel sluggish for production work where speed matters.”

Accuracy (40%)
85.0
Reliability (30%)
100.0
Latency (20%)
96.7
Cost Efficiency (10%)
100.0

Test Examples

Real requests we sent and the responses we received.

Search request in non-English language

POST /perplexity/searchedge1459ms
100%

{"data":{"id":"6633ac82-d0ee-49fb-8efe-18b491b9beff","results":[{"url":"https://es.wikipedia.org/wiki/Capitalidad_en_Espa%C3%B1a","date":"2008-03-23","title":"Capitalidad en España - Wikipedia, la enc

Search request with simple query

POST /perplexity/searchtypical1571ms
100%

{"success":true,"data":{"id":"6756c146-5767-4950-a304-dbea324c6aed","results":[{"date":"2024-12-20","last_updated":"2026-03-21","snippet":"Join us as we count down the most groundbreaking AI innovatio

Basic chat request with simple question

POST /perplexity/chattypical2882ms
100%

{"data":{"id":"b2209c99-ffa5-4722-aa20-3a83d3cd6f0c","model":"sonar","usage":{"cost":{"total_cost":0.00501,"request_cost":0.005,"input_tokens_cost":0.00001,"output_tokens_cost":0},"total_tokens":13,"p

Embedding request for text representation

POST /perplexity/embedtypical1328ms
20%

HTTP 502

Service Details

API AccessOpenAPI3rd Party
by Perplexity
POST /perplexity/chat
POST /perplexity/search
POST /perplexity/embed
POST /perplexity/context-embed
Homepage
Endpoint: https://perplexity.mpp.paywithlocus.com
Status: active
Discovered: 3/19/2026
Tests run: 10
Last tested: 3/21/2026

Test History

3/21/2026, 2:59:23 PM
score: 85%p50: 1949ms
3/21/2026, 2:27:49 PM
score: 3%p50: 235ms
3/20/2026, 7:47:34 PM
score: 4%p50: 241ms

Embed this badge

MPPrimo rating for Perplexity<a href="https://mpprimo.com/service/1d563376-c146-40b0-8b40-e540cc820530"><img src="https://mpprimo.com/api/badge/1d563376-c146-40b0-8b40-e540cc820530" alt="MPPrimo rating"></a>
How do we test? Read our methodology.