Seed 2.0 Pro (Non-reasoning): benchmark results
Provider: ByteDance. Released 2026-02-14. Access: API.
Unified ELO 1606 ± 1, rank #411 of 1935 rated models, from 27 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RuleWeaver - Cross-Source - Rule Precision | 75.73 | Correctly applied share of cited rules (%) | 90 |
| EgoBench - Dynamic Hard Mode - Micro Accuracy | 37.22 | MicroAcc (%; share of ground-truth tool calls issued, pooled | 85.7 |
| EgoBench - Static Mode - Micro Accuracy | 40.7 | MicroAcc (%; share of ground-truth tool calls issued, pooled | 85.7 |
| VitaBench 2.0 | 42.8 | Avg@4 Full Context (self-reported) | 75 |
| EgoBench - Dynamic Easy Mode - Micro Accuracy | 38.69 | MicroAcc (%; share of ground-truth tool calls issued, pooled | 71.4 |
| EgoBench - Static Mode | 13.3 | Joint Success Rate (%; every required tool call issued and t | 71.4 |
| ComboShoppingBench - Response Quality | 88.7 | Pass rate (%; LLM-judged) | 59.5 |
| EgoBench - Dynamic Easy Mode | 14.74 | Joint Success Rate (%; every required tool call issued and t | 57.1 |
| EgoBench - Dynamic Hard Mode | 11.58 | Joint Success Rate (%; every required tool call issued and t | 57.1 |
| RuleWeaver - Cross-Source - Rubric Score | 33.02 | Judge rubric score (0-100) | 50 |
| RuleWeaver - Same-Source - Rule Precision | 61.63 | Correctly applied share of cited rules (%) | 40 |
| LLMEval-Logic Formalization Free | 30.9 | Accuracy (%) | 38.5 |
Interactive version: theaggregate.ai/model?slug=seed-2-0-pro-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-25.