deepseek-math-7B-base: benchmark results
Provider: DeepSeek. Access: Open.
Aggregate Rating 1344 ± 39, rank #7851 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 59.21 | Accuracy (%) (5-shot) | 70.9 |
| Open LLM Leaderboard v1 - MMLU | 57.25 | Accuracy (%) (5-shot) | 40.7 |
| Open LLM Leaderboard v1 - ARC Challenge | 52.22 | Normalized accuracy (%) (25-shot) | 26.8 |
| Open LLM Leaderboard v1 - WinoGrande | 66.77 | Accuracy (%) (5-shot) | 20.6 |
| Open LLM Leaderboard v1 - HellaSwag | 69.49 | Normalized accuracy (%) (10-shot) | 20.1 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 40.71 | MC2 (%) (0-shot) | 15 |
Interactive version: theaggregate.ai/model?slug=deepseek-math-7b-base · How It Works · Data refreshed daily, snapshot 2026-11-08.