ERNIE-X1-Turbo-32K: benchmark results
Baidu's ERNIE X1 Turbo (April 25, 2025), a faster, lower-cost deep-thinking model with text and image understanding and a 32K context. Provider: Baidu. Released 2025-04-25. Access: API.
Aggregate Rating 1672 ± 23, rank #259 of 1635 rated models, from 131 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass Agent - Easy Multi-Turn (CompassBench 2507) | 76.5 | Score (%) | 100 |
| OpenCompass Knowledge - Science - English (CompassBench 2504) | 89 | Score (%) | 100 |
| OpenCompass LLM - Instruction Following (CompassBench 2504) | 71.5 | Score (%) | 100 |
| OpenCompass LLM - Instruction Following - Chinese (CompassBench 2504) | 76.2 | Score (%) | 100 |
| OpenCompass LLM - Language - English (CompassBench 2504) | 75.3 | Score (%) | 100 |
| OpenCompass LLM - Reasoning - Chinese (CompassBench 2504) | 92.8 | Score (%) | 100 |
| OpenCompass Language - Dialogue - English (CompassBench 2504) | 75.2 | Score (%) | 100 |
| OpenCompass Reasoning - Academic - Chinese (CompassBench 2504) | 93 | Score (%) | 100 |
| OpenCompass Reasoning - Puzzle - Chinese (CompassBench 2504) | 93.7 | Score (%) | 100 |
| OpenCompass Knowledge - Social Science - English (CompassBench 2504) | 94 | Score (%) | 98.1 |
| OpenCompass LLM - Reasoning (CompassBench 2504) | 92.3 | Score (%) | 96.3 |
| OpenCompass Language - Dialogue - Chinese (CompassBench 2504) | 94 | Score (%) | 96.3 |
Interactive version: theaggregate.ai/model?slug=ernie-x1-turbo-32k · How It Works · Data refreshed daily, snapshot 2026-10-15.