DeepSeek V3.2 (High): benchmark results
DeepSeek V3.2 evaluated at the high reasoning-effort setting. Provider: DeepSeek. Released 2025-12-01. Access: Open.
Unified ELO 1650 ± 1, rank #346 of 3078 rated models, from 37 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Horangi 4 - Korean Hate Speech | 70 | Accuracy (%) | 94.7 |
| Horangi 4 - HRM8K | 95.79 | Accuracy (%) | 88.5 |
| Horangi 4 - GLP - Mathematical Reasoning | 95.89 | Score (%) | 87.5 |
| Horangi 4 - Ko-AIME 2025 | 96 | Accuracy (%) | 81.7 |
| Horangi 4 - ALT Average | 77.53 | Score (%) | 79.3 |
| Horangi 4 - GLP - Specialized Knowledge | 53.28 | Score (%) | 76.9 |
| Horangi 4 - Ko-HLE | 26.76 | Accuracy (%) | 76 |
| Horangi 4 - BFCL | 68.22 | Accuracy (%) | 74.4 |
| Horangi 4 - KMMLU-Pro | 79.8 | Accuracy (%) | 74 |
| Horangi 4 - KoBALT-700 (Syntax) | 73 | Accuracy (%) | 73.6 |
| Horangi 4 - Ko-ARC-AGI | 57.69 | Accuracy (%) | 73.1 |
| SWE-bench Verified | 70 | Resolved (%) | 71.7 |
Interactive version: theaggregate.ai/model?slug=deepseek-v3-2-high · How It Works · Data refreshed daily, snapshot 2026-09-19.