Sonar Reasoning Pro: benchmark results
Provider: Perplexity. Access: API.
Aggregate Rating 1661 ± 21, rank #295 of 1635 rated models, from 17 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Tinybird AI SQL Benchmark - First-Attempt Success Rate | 98 | Questions answered with a valid query on the first attempt ( | 80.9 |
| Gapminder AI Worldview | 86.8 | Correct Rate (%) | 79.4 |
| CiteTrace - Fidelity Failure Rate | 18 | Fidelity failure rate (%; share of evaluable citations whose | 77.8 |
| AEPC-QA (RAG) | 66.79 | Accuracy (%) with the authors' retrieval-augmented pipeline | 72.1 |
| Tinybird AI SQL Benchmark - Success Rate | 100 | Questions answered with a valid query within 3 retries (%) | 71.9 |
| Tinybird AI SQL Benchmark - Exactness | 49.13 | Result exactness vs human reference queries (0-100) | 52.4 |
| AEPC-QA (Closed-Book) | 50.62 | Accuracy (%) answering from the model's own knowledge on AEP | 40 |
| Deep FinResearch Bench (Firm B Template) | 1.47 | Mean overall report-quality score (1-4 scale, 1 poor to 4 ex | 37.5 |
| Deep FinResearch Bench | 1.54 | Mean overall report-quality score (1-4 scale, 1 poor to 4 ex | 25 |
| CiteTrace - Suitability Failure Rate | 31.48 | Suitability failure rate (%; share of evaluable citations wh | 22.2 |
| DeepResearch Bench - Citation Accuracy | 45.19 | Score (%) | 16.7 |
| DeepResearch Bench - Effective Citations | 9.39 | Effective Citations | 16.7 |
Interactive version: theaggregate.ai/model?slug=sonar-reasoning-pro · How It Works · Data refreshed daily, snapshot 2026-11-05.