Perplexity AIproprietary
Released: 2025-02-05

Sonar Reasoning Pro

Perplexity premier reasoning model integrating deep test-time verification chains with live web citations and grounding.

Intelligence Index
90.2/ 100
Calibrated multi-domain composite
Throughput Speed
78tok/sec
Output streaming throughput
Time To First Token
350ms
Initial chunk server latency
LiveBench Contam-Free
80.5%
September 2026 suite score

LiveBench Multi-Domain Evaluation Breakdown

Monthly refreshed contamination-free test sets
Logical Reasoning87.0%
Coding & Repo Repair81.2%
Mathematics (AIME)86.0%
Data Analysis & Tables78.0%
Language & Comprehension75.5%
Instruction Following75.2%

Sourced Benchmark Evaluations (3)

Verified performance across authoritative benchmarks with provenance tracking.
BenchmarkStatusScoreTrustDateSource TypeProvenance
ARC-AGI-2Nearing Saturation70.2%
75
2025-03-05independentSource ↗
SWE-bench VerifiedSaturated64.2%
25
2025-02-14independentSource ↗
IFEvalDeprecated75.2%
25
2025-02-09independentSource ↗
OpenAI Compatible API ExecutionAPI Ready
from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="sonar-reasoning-pro",
    messages=[{"role": "user", "content": "Evaluate multi-step logic problem"}],
    temperature=0.2
)
print(response.choices[0].message.content)

Evaluation & Sourcing Notes

Scores listed for Sonar Reasoning Pro represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.

All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.