On 2026-10-28, the YZ Index Smoke quick test covered 13 models, with GPT-6 Luna ranking first for the day at 86.95 points. Smoke is a daily 10-question quick test, suitable for observing short-term signals, and is not equivalent to the conclusions of the Full weekly leaderboard.
This Smoke evaluation covered 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 long-term conclusions about model capability.
Daily Ranking
| Rank | Model | Main Leaderboard | Code Execution | Material Constraints | Integrity |
|---|---|---|---|---|---|
| #1 | GPT-6 Luna | 86.95 | 100 | 71 | pass |
| #2 | GPT-6 Astra | 83.94 | 100 | 64.3 | pass |
| #3 | GPT-6 Sol | 83.94 | 100 | 64.3 | pass |
| #4 | GPT-6.1 Sol | 83.94 | 100 | 64.3 | pass |
| #5 | Doubao Pro | 78.02 | 75 | 81.7 | pass |
| #6 | GPT-o3 | 78.02 | 75 | 81.7 | pass |
| #7 | Grok 4 | 74.67 | 74.4 | 75 | pass |
| #8 | GPT-5.5 | 73.2 | 75 | 71 | pass |
| #9 | Claude Opus 4.7 | 66.77 | 75 | 56.7 | pass |
| #10 | Claude Sonnet 4.6 | 66.66 | 74.8 | 56.7 | pass |
| #11 | Gemini 2.5 Pro | 56.44 | 50 | 64.3 | pass |
| #12 | DeepSeek V4 Pro | 45.7 | 25 | 71 | pass |
| #13 | Gemini 3.1 Pro | 45.7 | 25 | 71 | pass |
Main Changes
- Doubao Pro: Main leaderboard rose 32.7 points, code execution +50 points, material constraints +11.5 points
- Gemini 3.1 Pro: Main leaderboard fell 31.6 points, code execution -50 points, material constraints -9.2 points
- GPT-6 Sol: Main leaderboard rose 22 points, code execution +25 points, material constraints +18.4 points
- GPT-6.1 Sol: Main leaderboard rose 21 points, code execution +25 points, material constraints +16.2 points
- Claude Opus 4.7: Main leaderboard fell 20.9 points, code execution -24.8 points, material constraints -16.2 points
Signals to Watch
- Grok 4: Code execution plunged by 20.2 points
- GPT-5.5: Code execution plunged by 25 points
- Claude Opus 4.7: Main leaderboard plunged by 20.9 points
- Claude Sonnet 4.6: Main leaderboard plunged by 9.9 points
- DeepSeek V4 Pro: Main leaderboard plunged by 11.1 points
- Gemini 3.1 Pro: Main leaderboard plunged by 31.6 points
- GLM-4.6: Data incomplete (missing integrity and communication dimensions; API failure/timeout); automatic re-run has been initiated and it is not included in this period's ranking
- Qwen3 Max: Data incomplete (missing several dimensions, including execution, judgment, integrity, and communication; API failure/timeout); automatic re-run has been initiated and it is not included in this period's ranking
When reading this type of Smoke brief, focus on two questions: first, whether a model repeatedly shows the same type of weakness over multiple days; second, whether its integrity rating moves from pass to warn or fail. Large single-day changes in execution or constraint scores may come from question sampling or may be early signals of real degradation, and subsequent runs are needed for review.
Data source: YZ Index | Run #399 | View raw data
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接