Qwen 2.5 Coder 3B: benchmark results
Provider: Alibaba. Released 2024-11-12. Access: Open.
Aggregate Rating 1349 ± 31, rank #7726 of 10916 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| GSM8K | 75.7 | Accuracy (%) | 74.2 |
| C3-Bench (Controllable Code Completion) - Scale Control | 10.1 | Instruction-Following Rate (%; 909 SCC tasks: AST node-type | 57.5 |
| ARC Challenge (AI2) | 52.9 | Accuracy (%) | 43.4 |
| C3-Bench (Controllable Code Completion) - Implementation Control - Pass@1 | 44.8 | Pass@1 (%; unit tests on the 1,286 instructed ICC tasks) | 42.5 |
| C3-Bench (Controllable Code Completion) - Implementation Control | 26.3 | Instruction-Following Rate (%; 1,286 ICC tasks: passes the u | 37.5 |
| DomainCodeBench | 87.46 | Composite Score (%) | 33.3 |
| HellaSwag | 70.9 | Accuracy (%) | 32.9 |
| WinoGrande | 67.4 | Accuracy (%) | 31.2 |
| GeoCode Leaderboard | 45.05 | AutoGEEval++ Pass@1 (imported) | 27.3 |
| Epoch AI - ECI | 97.32 | ECI Score | 14.4 |
Interactive version: theaggregate.ai/model?slug=qwen-2-5-coder-3b · How It Works · Data refreshed daily, snapshot 2026-11-08.