DeepSeek V3.2 (Non-reasoning): benchmark results
DeepSeek V3.2 evaluated with reasoning disabled. Provider: DeepSeek. Released 2025-12-01. Access: Open.
Aggregate Rating 1611 ± 6, rank #679 of 2168 rated models, from 364 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ClarifyCodeBench - Key Question Rate (TKQR) | 0.3 | Turn-discounted key question rate (0 to 1; nDCG-style credit | 100 |
| ClarifyCodeBench - Optimal Round Adherence (ORA) | 0.5 | Optimal round adherence (0 to 1; Gaussian credit exp(-(n-K)^ | 100 |
| ConflictQA - Text Conflict (Non-Complementary) | 53.43 | Exact match (%) on 802 single-hop (non-complementary) questi | 100 |
| GR-Ben - Abductive Reasoning | 71.4 | F1 (%) of first-error identification: the harmonic mean of t | 100 |
| GR-Ben - Chemistry | 50.3 | F1 (%) of first-error identification: the harmonic mean of t | 100 |
| GR-Ben - Physics | 61.5 | F1 (%) of first-error identification: the harmonic mean of t | 100 |
| UnpredictaBench - Text | 28.13 | KS@100 (%): share of the 160 Text problems (named or describ | 97.4 |
| UGI Leaderboard | 55.75 | UGI Score | 96.4 |
| GR-Ben - Biology | 49.9 | F1 (%) of first-error identification: the harmonic mean of t | 95.2 |
| GR-Ben - Computer Science | 58.5 | F1 (%) of first-error identification: the harmonic mean of t | 95.2 |
| GR-Ben - Deductive Reasoning | 44.9 | F1 (%) of first-error identification: the harmonic mean of t | 95.2 |
| GR-Ben - Inductive Reasoning | 67.5 | F1 (%) of first-error identification: the harmonic mean of t | 95.2 |
Interactive version: theaggregate.ai/model?slug=deepseek-v3-2-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.