Directory · Models · Perplexity

Sonar

Sonar Large

Name: Sonar Large
Brand: Perplexity
Price: 1 USD

Llama-tuned with a search-first system prompt. The canonical answer-with-citations model.

Input / 1M

$1.00

Output / 1M

$5.00

Context

128K

Released

Oct 7, 2025

Benchmark profile

Profile

MMLU-Pro79.2%

HumanEval78.5%

IFEval84.6%

TruthfulQA84.6%

HarmBench84.6%

LiveBench79.2%

Scores

Editorial

Strengths

+Live web search built-in
+Citations by default

Trade-offs

−Limited creative range

Price history (last 90 days)

History starts accumulating on the next nightly snapshot.

What this costs to use

All task cards →

Summarize a 1-hour meeting transcript
15,000 in · 500 out
$0.017per task
Review a 500-line pull request
8,000 in · 2,000 out
$0.018per task
Answer a customer support ticket
3,000 in · 800 out
$0.0070per task

Compare across every model →

Recommended alternatives

Cheaper

Claude Haiku 4.5

Anthropic

$0.80/M · $4.00/M−20%

Stronger

Claude Sonnet 4.6

Anthropic

$3.00/M · $15.00/M+200%

Direct competitor

Jamba 2

AI21

$0.50/M · $1.50/M−67%

Where to use it