Seed 2.0 Pro (High): benchmark results
Provider: ByteDance. Released 2026-02-14. Access: API.
Aggregate Rating 1731 ± 12, rank #278 of 2168 rated models, from 101 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CLBench Life - Community Interactions | 23 | Task solving rate (%) on the 45 community-interaction (forum | 100 |
| LivingScreen - Cross-Source Understanding | 68.3 | Success rate (%) on L2 multiple-choice questions linking vid | 100 |
| SceneActBench | 50.2 | Overall Score (%) | 100 |
| SceneActBench - Camera | 34.5 | Task score (0-100): 100 furnished 3D-FRONT scenes: recover t | 100 |
| SceneActBench - Dynamic | 70.7 | Task score (0-100): 10 low-poly Kenney scenes from a sampled | 100 |
| WildTableBench - Semantic Lookup and Structure | 56.7 | Accuracy (%) on the Semantic Lookup and Structure subtype (C | 97.5 |
| MSQA - Chinese | 55.7 | F-score (%; harmonic mean of CO, the share of fully correct | 95 |
| WildTableBench - Aggregate Verification | 63.6 | Accuracy (%) on the Aggregate Verification subtype (C3-A), 9 | 95 |
| WildTableBench - Cell Locating | 67.3 | Accuracy (%) on the Cell Locating subtype (C1-L), 928 human- | 95 |
| WildTableBench - Cell-Level | 59.3 | Accuracy (%) on the Cell-Level question category (C1, questi | 95 |
| WildTableBench - Excel Formula | 58.3 | Accuracy (%) on the Excel Formula subtype (C1-F), 928 human- | 95 |
| WildTableBench - Transcription | 50 | Accuracy (%) on the Transcription subtype (C1-T), 928 human- | 95 |
Interactive version: theaggregate.ai/model?slug=seed-2-0-pro-high · How It Works · Data refreshed daily, snapshot 2026-10-20.