Qwen3-1.7B-Base: benchmark results
Alibaba's Qwen3-1.7B-Base, the pretrained (non-instruct) dense 1.7B checkpoint in the Qwen3 series, trained on 36T tokens across 119 languages with a three-stage pretraining pipeline. Provider: Alibaba. Released 2025-04-29. Access: Open.
Aggregate Rating 1411 ± 25, rank #1171 of 1635 rated models, from 164 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Polar (US) - Sociocultural Axis | 84.29 | Sociocultural-axis ICAT (0-100), the StereoSet idealized con | 79.7 |
| Polar (US) - Economic Axis | 71.22 | Economic-axis ICAT (0-100), the StereoSet idealized context | 77 |
| Polar (US) | 77.75 | Total ICAT (0-100), the StereoSet idealized context associat | 75.7 |
| Swallow - Pre-trained English - MATH | 45.6 | Exact Match (%) | 75.5 |
| Open Arabic LLM - Arabic MMLU Civics (High School) | 45.98 | Accuracy (%) | 68.5 |
| Open Arabic LLM - Arabic MMLU HT Electrical Engineering | 53.79 | Accuracy (%) | 67.6 |
| Swallow - Pre-trained English - HumanEval | 46.2 | Pass@1 (%) | 63.3 |
| Open Arabic LLM - Arabic MMLU HT High School Mathematics | 32.59 | Accuracy (%) | 62.3 |
| Open Arabic LLM - Arabic MMLU HT Virology | 45.78 | Accuracy (%) | 60.8 |
| Open Arabic LLM - Arabic MMLU HT Elementary Mathematics | 43.65 | Accuracy (%) | 59.9 |
| Open Arabic LLM - Arabic MMLU HT Abstract Algebra | 30 | Accuracy (%) | 59.3 |
| Open Arabic LLM - Arabic MMLU HT Machine Learning | 37.5 | Accuracy (%) | 57.4 |
Interactive version: theaggregate.ai/model?slug=qwen3-1-7b-base · How It Works · Data refreshed daily, snapshot 2026-10-15.