Seed 2.0 Pro (Medium): benchmark results
Provider: ByteDance. Released 2026-02-14. Access: API.
Aggregate Rating 1666 ± 20, rank #470 of 2170 rated models, from 31 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MSQA - Chinese | 50.2 | F-score (%; harmonic mean of CO, the share of fully correct | 80 |
| CLBench - Empirical Discovery & Simulation | 13.9 | Solving Rate (%) | 77.1 |
| CLBench - Rule System Application | 19.3 | Solving Rate (%) | 71.4 |
| CLBench | 18.7 | Solving Rate (%) | 58.6 |
| CLBench - Domain Knowledge Reasoning | 20.1 | Solving Rate (%) | 57.1 |
| MSQA - Beliefs, Values and Knowledge Systems | 45.2 | F-score (%; harmonic mean of CO, the share of fully correct | 55 |
| MSQA - Japanese | 25.9 | F-score (%; harmonic mean of CO, the share of fully correct | 55 |
| CLBench Life - Behavioral Records & Activity Trails | 11.9 | Solving Rate (%) | 53.6 |
| PlanningBench - Avg-pass | 77.09 | Average checklist pass rate (%): share of an instance's veri | 52.9 |
| MSQA | 41.2 | F-score (%; harmonic mean of CO, the share of fully correct | 50 |
| MSQA - Korean | 32.6 | F-score (%; harmonic mean of CO, the share of fully correct | 50 |
| MSQA - History and Collective Memory | 39.7 | F-score (%; harmonic mean of CO, the share of fully correct | 45 |
Interactive version: theaggregate.ai/model?slug=seed-2-0-pro-medium · How It Works · Data refreshed daily, snapshot 2026-11-05.