Sabiazinho-3: benchmark results
Provider: Other.
Aggregate Rating 1561 ± 48, rank #1537 of 10916 rated models, from 9 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CAPITU - Easy | 94.3 | Strict prompt-level accuracy (%) on the 70 easy prompts (one | 41.2 |
| CAPITU - Multi-Turn | 68 | Conversation accuracy (%): share of the 100 three-turn CAPIT | 29.4 |
| CAPITU - Hard | 60 | Strict prompt-level accuracy (%) on the 60 hard prompts (thr | 23.5 |
| CAPITU - Strict Accuracy | 80 | Strict prompt-level accuracy (%): share of prompts whose con | 23.5 |
| Magis-Bench | 3.14 | AVG (self-reported) | 22.7 |
| CAPITU - Instruction-Level Accuracy | 87.4 | Instruction-level accuracy (%): share of all individual cons | 17.6 |
| CAPITU - Medium | 82.9 | Strict prompt-level accuracy (%) on the 70 medium prompts (t | 17.6 |
| CAPITU | 90.2 | Final score (%): the mean over prompts of (2 x the share of | 11.8 |
| CAPITU - Coherence | 91.7 | Coherence score (%): a GPT-4o-mini judge's 0-1 rating of the | 0 |
Interactive version: theaggregate.ai/model?slug=sabiazinho-3 · How It Works · Data refreshed daily, snapshot 2026-11-08.