Qwen2.5-Coder-1.5B-Instruct: benchmark results
Alibaba's Qwen2.5-Coder-1.5B-Instruct, the instruction-tuned version of Qwen2.5-Coder-1.5B, part of the six-size Qwen2.5-Coder family pretrained on 5.5 trillion tokens with a 32,768-token context. Provider: Alibaba. Released 2024-09-19. Access: Open.
Aggregate Rating 1309 ± 19, rank #1460 of 1656 rated models, from 40 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| C3-Bench (Controllable Code Completion) - Scale Control | 8 | Instruction-Following Rate (%; 909 SCC tasks: AST node-type | 47.5 |
| MERA - BPS | 94.6 | Accuracy (%) | 46 |
| Delulu - Edit Similarity | 63.2 | Edit similarity (0-100): character-level normalized Levensht | 40 |
| MERA - RWSD | 52.69 | Accuracy (%) | 34.5 |
| Delulu | 0.65 | pass@1 | 30 |
| Delulu - Exact Match | 37.8 | Exact match (%) of the completion with the gold completion, | 30 |
| MERA - ruHumanEval | 2.01 | pass@1 (%) | 28.7 |
| MERA - ruCodeEval | 0.73 | pass@1 (%) | 28.3 |
| MERA - SimpleAr | 92.9 | EM (%) | 28.1 |
| MERA - LCS | 10.2 | Accuracy (%) | 27.9 |
| MERA - ruTiE | 57.31 | Accuracy (%) | 25.7 |
| MERA - MultiQ | 23.6 | F1 (%) | 24.2 |
Interactive version: theaggregate.ai/model?slug=qwen2-5-coder-1-5b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.