The 2026-10-20 YZ Index Smoke quick test covered 14 models, with GPT-o3 ranking first for the day at 97.83. Smoke is a daily 10-question quick test suitable for observing short-term signals and is not equivalent to conclusions from the weekly Full leaderboard.
This Smoke evaluation covers only two main leaderboard dimensions: code execution and material constraints. The main leaderboard formula is 0.55 × code execution + 0.45 × material constraints. Because the daily sample size is small, single-day scores are better used as monitoring signals rather than as long-term conclusions about model capability.
Daily Ranking
| Rank | Model | Main | Code execution | Material constraints | Integrity |
|---|---|---|---|---|---|
| #1 | GPT-o3 | 97.83 | 98.5 | 97 | pass |
| #2 | Claude Opus 4.7 | 91.2 | 97 | 84.1 | pass |
| #3 | GPT-5.5 | 91.2 | 97 | 84.1 | pass |
| #4 | Doubao Pro | 90.48 | 95.7 | 84.1 | pass |
| #5 | GPT-6 Luna | 89.04 | 97 | 79.3 | pass |
| #6 | Grok 4 | 88.89 | 92.8 | 84.1 | warn |
| #7 | Claude Sonnet 4.6 | 85.71 | 97 | 71.9 | pass |
| #8 | GPT-6.1 Sol | 84.76 | 97 | 69.8 | pass |
| #9 | GPT-6 Sol | 82.6 | 97 | 65 | pass |
| #10 | Gemini 2.5 Pro | 72.75 | 75 | 70 | pass |
| #11 | GPT-6 Astra | 71.01 | 72 | 69.8 | pass |
| #12 | Gemini 3.1 Pro | 63.7 | 47 | 84.1 | pass |
| #13 | GLM-4.6 | 63.7 | 47 | 84.1 | pass |
| #14 | DeepSeek V4 Pro | 59.6 | 47 | 75 | pass |
Key Changes
- GPT-o3: main leaderboard up 42.7 points, code execution +51.5, material constraints +31.9
- Doubao Pro: main leaderboard up 16 points, code execution +23.7, material constraints +6.5
- GLM-4.6: main leaderboard up 15.4 points, material constraints +34.1
- Claude Opus 4.7: main leaderboard up 11.4 points, code execution +25, material constraints -5.2
- Claude Sonnet 4.6: main leaderboard up 11.4 points, code execution +25.4, material constraints -5.7
Signals to Watch
- GPT-6 Astra: main leaderboard plunges -11.3 points
- Qwen3 Max: data incomplete (several evaluation dimensions missing, API failure/timeout), automatically queued for a rerun, and not included in this ranking
When reading this kind of Smoke brief, focus on two questions: first, whether a model has exposed the same type of weakness for several consecutive days; second, whether its integrity rating has moved from pass to warn or fail. Large single-day changes in execution or constraint scores may come from question sampling or may be an early signal of real degradation and require follow-up runs for review.
Data source: YZ Index | Run #387 | View raw data
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接