qwen3.8-max-internal: benchmark results
Vals evaluation endpoint qwen3.8-max-internal. The exact internal endpoint is distinct from public Qwen3.8 Max; public API access or downloadable weights have not been established. Provider: Alibaba.
Aggregate Rating 1863 ± 37, rank #30 of 10916 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Vals AI Tax Agent Bench | 49.66 | Accuracy (%) | 98.8 |
| Vals AI Excel Modeling | 71.98 | Accuracy (%) | 89.5 |
| Vals AI Finance Index | 61.66 | Finance Index (%) | 86.1 |
| Vals AI Legal Index | 73.9 | Legal Index (%) | 85.3 |
| Vals AI Finance Agent v2 | 57.87 | Accuracy (%) | 85.2 |
| Vals AI Legal Research Bench | 46.15 | Accuracy (%) | 80.6 |
| Vals AI Vibe Code Bench 1-100 | 24.83 | Accuracy (%) | 76 |
| Vals AI Terminal-Bench 4.0 | 38.89 | Accuracy (%) | 74.2 |
| Vals AI Excel Modeling v1.1 | 71.98 | Accuracy (%) | 73.7 |
| Vals AI Harvey Legal Agent Bench v1.1 | 15.42 | Task Resolution (%) | 73.5 |
| Vals AI Vibe Code Bench | 77.14 | Accuracy (%) | 69.4 |
| Vals AI Code Migration v1.1 | 57.19 | Hidden-test Pass Rate (%) | 67.9 |
Interactive version: theaggregate.ai/model?slug=qwen3-8-max-internal · How It Works · Data refreshed daily, snapshot 2026-11-08.