Qwen3.8 Max Preview
qwen
Preview Qwen flagship for million-token multimodal reasoning and long-horizon agentic workflows
ReasoningTool CallingAttachments
Context Window
1M
Max Output
131K
Temperature
Yes
Open Weights
No
Knowledge Cutoff
N/A
Released
2026-07-19
Last Updated
2026-07-19
Modalities
Input:TextImageVideo
→Output:Text
Benchmarks
| Benchmark | Score | Metric | Harness |
|---|---|---|---|
| Terminal-Bench | 86.6 | accuracy | - |
| SWE-Bench Pro | 67.7 | resolve rate | Claude Code |
| DeepSWE | 56.6 | resolve rate | Claude Code |
| NL2Repo | 55.9 | resolve rate | Claude Code |
| FrontierSWE | 73.5 | dominance score | Claude Code |
| MLS-Bench-Lite | 41 | score | Claude Code |
| AutomationBench | 27.3 | pass@1 | - |
| Toolathlon Verified | 72.5 | pass@1 | - |
| WideSearch | 81.9 | F1 | - |
| Humanity's Last Exam | 56.2 | accuracy | - |
| GPQA Diamond | 92.6 | accuracy | - |
| Humanity's Last Exam | 43.6 | accuracy | - |
| IFBench | 82.8 | score | - |
| OSWorld-Verified | 86.1 | success rate | - |
| MMMU Pro | 82.3 | accuracy | - |
Available Providers (0)
Tidak ada provider yang terdaftar untuk model ini.