Sonar Pro Search: benchmark results
Provider: Perplexity. Access: API.
Aggregate Rating 1589 ± 52, rank #1038 of 10800 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SnakeBench | 25.7 | TrueSkill Rating | 69.6 |
| GISA - Set EM | 20 | Exact match (%) on the 50 set-type queries (an unordered col | 26.7 |
| GISA - Table Row F1 | 34.76 | Row-level F1 (%) between predicted and gold table rows on th | 20 |
| GISA - Item EM | 22.73 | Exact match (%) on the 22 item-type queries (a single answer | 16.7 |
| GISA - Set F1 | 47.04 | F1 (%) between the predicted and gold sets on the 50 set-typ | 13.3 |
| GISA - Table Item F1 | 49.05 | Item-level F1 (%) between predicted and gold table cells on | 13.3 |
| Tinybird AI SQL Benchmark - Exactness | 34.95 | Result exactness vs human reference queries (0-100) | 12 |
| GISA - Table EM | 3.95 | Exact match (%) on the 253 table-type queries (a table with | 6.7 |
| Tinybird AI SQL Benchmark - Success Rate | 66 | Questions answered with a valid query within 3 retries (%) | 6.3 |
| Tinybird AI SQL Benchmark - First-Attempt Success Rate | 0 | Questions answered with a valid query on the first attempt ( | 0.1 |
| GISA - List EM | 6.25 | Exact match (%) on the 48 list-type queries (an ordered list | 0 |
| GISA - List F1 | 34.74 | F1 (%) of list items regardless of order on the 48 list-type | 0 |
Interactive version: theaggregate.ai/model?slug=sonar-pro-search · How It Works · Data refreshed daily, snapshot 2026-11-13.