Qwen 3.8 Max
Qwen
qwen/qwen3.8-max-0902
Rank likelihood
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
|---|---|---|---|---|---|---|---|---|---|---|
| 2% | 5% | 5% | 6% | 11% | 6% | 7% | 13% | 13% | 13% | 18% |
Head-to-head results
| Opponent | 0%25%50%75%100% | Battles | Win rate |
|---|---|---|---|
Muse WonderMeta |
43 | 63% | |
Gemini 4 ArgonGoogle |
38 | 57% | |
DeepSeek V4.1 Flash 1017DeepSeek |
33 | 56% | |
Kimi K3.1 ProMoonshot AI |
19 | 47% | |
GPT-6 AstraOpenAI |
13 | 46% | |
Grok 4.7xAI |
32 | 44% | |
| GPT-6.1 SolOpenAI |
22 | 41% | |
MiMo-V2.6 FlashXiaomi |
29 | 40% | |
Claude Opus 5.5Anthropic |
37 | 39% | |
| Claude Fable 5.1Anthropic |
39 | 33% |
Elo by week
Last week based on partial data
Details
| Eligible battles | 348 |
|---|---|
| Unique voters | 24 |
| Raw win rate | 44% |
| Tokens per second | 30.9 |
| Lab | Qwen |
| Served through | OpenRouter |
| Model ID | qwen/qwen3.8-max-0902 |
| Version | qwen/qwen3.8-max-20260902 |
| Thinking | Enabled · high (default) |
| First eligible vote | Oct 2, 2026 |