DeepSeek R1 Distill Qwen 32B: benchmark results
DeepSeek's official R1 reasoning distill onto Qwen2.5-32B (MIT license), SFT-trained on ~800K R1-generated reasoning traces. Provider: DeepSeek. Released 2025-01-20. Access: Open.
Aggregate Rating 1537 ± 6, rank #655 of 1656 rated models, from 692 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EPPCMinerBen (0-Shot) - Subcode Classification | 48.25 | Micro-averaged F1 (%) of EPPC subcode classification (choosi | 100 |
| EPPCMinerBen (1-Shot) - Subcode Classification | 47.87 | Micro-averaged F1 (%) of EPPC subcode classification (choosi | 100 |
| EPPCMinerBen (2-Shot) - Subcode Classification | 47.95 | Micro-averaged F1 (%) of EPPC subcode classification (choosi | 100 |
| Open CoT - LogiQA 2 | 19.02 | CoT Gain (%) | 100 |
| PARROT - MySQL Dialect Compatibility | 58.82 | Accuracy (%) | 100 |
| PERSPECTRA - Opinion Matching | 85.4 | Accuracy (%) of identifying which of the topic's pro and con | 100 |
| PERSPECTRA - Opinion Matching Stance Accuracy | 95 | Stance accuracy (%): the chosen candidate opinion is on the | 100 |
| RFEval - Legal Decision | 91.04 | Contrast-conditional reasoning faithfulness (%): after a cou | 100 |
| MERA - PARus | 96.2 | Accuracy (%) | 99.5 |
| Open CoT Leaderboard | 16.92 | Average CoT Gain (%) | 99.2 |
| Thai LLM - Extraction | 7.85 | Rating (0-10) | 98.6 |
| Open CoT - LSAT Analytical Reasoning | 16.52 | CoT Gain (%) | 98.5 |
Interactive version: theaggregate.ai/model?slug=deepseek-r1-distill-qwen-32b · How It Works · Data refreshed daily, snapshot 2026-10-20.