DeepSeek V3.2 (Thinking): benchmark results
DeepSeek V3.2 evaluated with thinking enabled. Provider: DeepSeek. Released 2025-12-01. Access: Open.
Unified ELO 1617 ± 1, rank #362 of 1761 rated models, from 136 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SuperCLUE-Mkt - Compliance Assessment | 92.34 | Score | 100 |
| SuperCLUE-Mkt - Placement Strategy | 89.66 | Score | 100 |
| UGI Leaderboard | 57.78 | UGI Score | 98.1 |
| UGI - Writing | 57.28 | Writing Score | 91.6 |
| SuperCLUE-Mkt - Market Insight | 88.12 | Score | 90.9 |
| SuperCLUE-Mkt - Overall | 85.74 | Score | 90.9 |
| UGI - Natural Intelligence | 48.11 | NatInt Score | 89.7 |
| C4STYLI | 74.5 | Accuracy (C4Styli-T) (self-reported) | 88.2 |
| LLM2014 Logic 2025-10 | 54.16 | Median Score | 87.8 |
| BenchTable | 70.7 | Total Score (%) | 87.5 |
| AA TAU-2 Bench | 90.64 | Accuracy (%) | 85.4 |
| LLM2014 Code 2025-11 - TypeScript | 8.17 | Score | 84 |
Interactive version: theaggregate.ai/model?slug=deepseek-v3-2-thinking · How It Works · Data refreshed daily, snapshot 2026-09-05.