Qwen 3.6 Plus: benchmark results
Alibaba Qwen 3.6 Plus model for high-end chat, coding, and reasoning workloads. Provider: Alibaba. Released 2026-04-02. Access: API.
Aggregate Rating 1750 ± 3, rank #120 of 1656 rated models, from 1306 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AARRI-Bench (Hermes Agent) - Hands-on | 64.3 | Task success rate (%) under the 0/1 reward on the hands-on t | 100 |
| AARRI-Bench (Mini-SWE-Agent) - Hands-on | 64.3 | Task success rate (%) under the 0/1 reward on the hands-on t | 100 |
| ArabicDialectSafety - Unsafe Response Rate | 0.2 | Unsafe response rate (%; share of 500 stratified harmful dia | 100 |
| BIM-Edit - Topology | 47.77 | Topology score (0-100): weighted F1 of edited nodes (0.3) an | 100 |
| CC-OCR V2 | 75.77 | Average (self-reported) | 100 |
| CUDABeaver - Pass@1 | 15.5 | Share of the 213 broken-start CUDA debugging tasks (%) whose | 100 |
| CalBench - Uniform Costs - Coordination | 93.33 | Coordination (%) | 100 |
| From Knowing to Doing | 85.29 | Total ret. (self-reported) | 100 |
| HIDBench - NodLink HW20 | 73.9 | Matthews correlation coefficient of malicious-event detectio | 100 |
| HPR-Bench - Tag Inference | 73.29 | Avg@10 Tag Accuracy (%; mean of 10 samples over 18 profile t | 100 |
| HPR-Bench - Tag Inference - Geographic Context | 79.93 | Avg@10 Tag Accuracy (%; mean of 10 samples) | 100 |
| HPR-Bench - Tag Inference - Life Stage | 79.07 | Avg@10 Tag Accuracy (%; mean of 10 samples) | 100 |
Interactive version: theaggregate.ai/model?slug=qwen-3-6-plus · How It Works · Data refreshed daily, snapshot 2026-10-20.