Qwen 3.8 Max (0902) (xHigh): benchmark results
Provider: Alibaba. Released 2026-09-02. Access: API.
Aggregate Rating 1800 ± 25, rank #137 of 2170 rated models, from 12 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OTIS Mock AIME 2024-25 | 100 | Accuracy (%) | 98.1 |
| Epoch AI - GPQA Diamond | 92.3 | Accuracy (%) | 91.3 |
| ObviousBench | 98.61 | Answer pass³ (%) | 83.2 |
| DuelLab GameBench 2 - Reasoning Variants | 68.12 | GameBench2 normalized score (0–100) | 82.2 |
| FrontierMath - Tiers 1-3 (v2) | 65.61 | Accuracy (%, 285 private v2 problems) | 73.9 |
| Chess Puzzles (Epoch AI) | 37 | Accuracy (%) | 70.9 |
| SimpleQA Verified | 47.3 | Accuracy (%) | 65.7 |
| RSI-Exam | 0.39 | Mean hidden-set score | 61.1 |
| FrontierMath - Tier 4 (v2) | 34.15 | Accuracy (%, 41 private v2 problems) | 58.9 |
| Paydirt - Held Pair Accuracy | 66.67 | Correct pairs (%); both answers must pass, raw arm, omp harn | 38.1 |
| FrontierSWE V2 | 17.81 | Mean@5 Score (%) | 18.2 |
| Epoch AI - Furniture Assembly | 20 | Score | 0 |
Interactive version: theaggregate.ai/model?slug=qwen-3-8-max-0902-xhigh · How It Works · Data refreshed daily, snapshot 2026-11-05.