Qwen 3 8B (Non-reasoning): benchmark results
Qwen 3 8B evaluated with reasoning disabled. Provider: Alibaba. Released 2025-04-29. Access: Open.
Aggregate Rating 1460 ± 6, rank #1233 of 2170 rated models, from 657 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BAGEL - Xeno-canto Modulation | 38.2 | Accuracy (%, times 100) on the 419 Xeno-canto questions on m | 100 |
| SRD Reference - AIME 2024 (non-thinking) | 26.94 | avg@12 (%) | 100 |
| SRD Reference - AIME 2025 (non-thinking) | 21.39 | avg@12 (%) | 100 |
| SRD Reference - AIME 2026 (non-thinking) | 15.83 | avg@12 (%) | 100 |
| SRD Reference - HMMT 2025 (non-thinking) | 11.53 | avg@12 (%) | 100 |
| SRD Reference - HMMT 2026 (non-thinking) | 20.71 | avg@12 (%) | 100 |
| SRD Reference - IFEval (non-thinking) | 81.7 | Strict prompt-level accuracy (%) | 100 |
| SRD Reference - MMLU-Pro (non-thinking) | 64.98 | Accuracy (%) | 100 |
| EuroEval Ukrainian Summarization - LR SUM UK | 31.21 | chrF3++ (%) | 93.8 |
| EuroEval Italian Summarization - Ilpost SUM | 38.55 | chrF3++ (%) | 93.3 |
| EuroEval Catalan Summarization - Dacsa CA | 38.65 | chrF3++ (%) | 92.2 |
| TRACE Bench - Length | 98.32 | Reply-length compliance (%) | 92 |
Interactive version: theaggregate.ai/model?slug=qwen-3-8b-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.