Doubao-Seed-1.6-Flash: benchmark results
Provider: ByteDance. Access: Open.
Unified ELO 1559 ± 23, rank #576 of 1629 rated models, from 39 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AGI-Eval Community - Cognition (English) | 84.94 | Accuracy (%) | 86.2 |
| AGI-Eval Community - Subject Knowledge | 86.47 | Accuracy (%) | 73.6 |
| AEGIS (Academic Image Forensics) | 36.03 | Normalized Forensic Index (%): harmonic mean of the macro-F1 | 56 |
| AGI-Eval Community - Learning (Chinese) | 78.8 | Accuracy (%) | 54.3 |
| AEGIS (Academic Image Forensics) - Textual Artifact Recognition | 60.39 | Macro-F1 (%) over two classes, telling whether the text regi | 44 |
| SpecVQA - English Reasoning (L1) | 56.87 | Accuracy (0-1, times 100) on the English reasoning (L1) ques | 42.1 |
| SpecVQA | 58.67 | Accuracy (0-1, times 100), the unweighted mean of the Englis | 36.8 |
| SpecVQA - Chinese Descriptive (L0) | 61.48 | Accuracy (0-1, times 100) on the Chinese descriptive (L0) qu | 36.8 |
| SpecVQA - Chinese Reasoning (L1) | 55.74 | Accuracy (0-1, times 100) on the Chinese reasoning (L1) ques | 36.8 |
| SpecVQA - English Descriptive (L0) | 60.6 | Accuracy (0-1, times 100) on the English descriptive (L0) qu | 36.8 |
| AGI-Eval Community - Learning | 81.34 | Accuracy (%) | 31.4 |
| AGI-Eval Community - Cognition | 73.63 | Accuracy (%) | 30.7 |
Interactive version: theaggregate.ai/model?slug=doubao-seed-1-6-flash · How It Works · Data refreshed daily, snapshot 2026-10-07.