The 2026-11-19 YZ Index Smoke Quick Test covered 13 models, with GPT-6 Sol ranking first for the day at 80.54 points. Smoke is a daily 10-question quick test, suitable for observing short-term signals and not equivalent to Full weekly leaderboard conclusions.
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 as long-term conclusions about model capability.
Daily Ranking
| Ranking | Model | Main Leaderboard | Code Execution | Material Constraints | Integrity |
|---|---|---|---|---|---|
| #1 | GPT-6 Sol | 80.54 | 68.7 | 95 | pass |
| #2 | GPT-5.5 | 78.62 | 74.3 | 83.9 | pass |
| #3 | GPT-o3 | 78.08 | 73.8 | 83.3 | pass |
| #4 | Claude Sonnet 4.6 | 77.41 | 68.5 | 88.3 | pass |
| #5 | Claude Opus 4.7 | 74.43 | 57.6 | 95 | pass |
| #6 | Doubao Pro | 71.84 | 48.8 | 100 | pass |
| #7 | Gemini 3.1 Pro | 68.77 | 47.3 | 95 | pass |
| #8 | GPT-6 Astra | 64.87 | 49.3 | 83.9 | pass |
| #9 | GPT-6.1 Sol | 64.87 | 49.3 | 83.9 | pass |
| #10 | GPT-6 Luna | 63.76 | 38.2 | 95 | pass |
| #11 | Grok 4 | 50.41 | 23.5 | 83.3 | pass |
| #12 | DeepSeek V4 Pro | 46.94 | 16.7 | 83.9 | pass |
| #13 | Gemini 2.5 Pro | 37.76 | 0 | 83.9 | pass |
Data Interpretation
Today's top five models on the main leaderboard show different balances between Code Execution and Material Constraints. GPT-6 Sol scored 68.7 in Code Execution and 95 in Material Constraints for a main leaderboard score of 80.54; GPT-5.5 scored 74.3 in Code Execution and 83.9 in Material Constraints for 78.62; GPT-o3 scored 73.8 in Code Execution and 83.3 in Material Constraints for 78.08; Claude Sonnet 4.6 scored 68.5 in Code Execution and 88.3 in Material Constraints for 77.41; Claude Opus 4.7 scored 57.6 in Code Execution and 95 in Material Constraints for 74.43. Among the leading models, the combination with 95 in Material Constraints still maintained a relatively high position under the main leaderboard formula, while the model with 74.3 in Code Execution narrowed the gap through its execution score.
In terms of anomalous signals, Gemini 2.5 Pro fell 44.8 points on the main leaderboard, 75 points in Code Execution, and 7.8 points in Material Constraints; DeepSeek V4 Pro fell 34.5 points on the main leaderboard, 58.3 points in Code Execution, and 5.5 points in Material Constraints; Grok 4 fell 33.5 points on the main leaderboard, 55.9 points in Code Execution, and 6.1 points in Material Constraints; GPT-6 Astra fell 26.3 points on the main leaderboard and 50.7 points in Code Execution; GPT-6.1 Sol fell 25.3 points on the main leaderboard and 50.7 points in Code Execution while rising 5.8 points in Material Constraints; GPT-o3 fell 16.7 points in Material Constraints; Doubao Pro fell 26.2 points in Code Execution; GPT-6 Luna fell 20.2 points on the main leaderboard. These large single-day fluctuations may stem from question sampling variation or may be signs of genuine degradation, and follow-up review is needed to confirm.
GLM-4.6 and Qwen3 Max had incomplete data due to API failures and did not participate in the current ranking. The Smoke Quick Test is a small-sample, single-day signal; the observations above reflect only the score structure for that day, and no long-term judgment is made for now.
Key Changes
- Gemini 2.5 Pro: Main leaderboard down 44.8 points, Code Execution -75 points, Material Constraints -7.8 points
- DeepSeek V4 Pro: Main leaderboard down 34.5 points, Code Execution -58.3 points, Material Constraints -5.5 points
- Grok 4: Main leaderboard down 33.5 points, Code Execution -55.9 points, Material Constraints -6.1 points
- GPT-6 Astra: Main leaderboard down 26.3 points, Code Execution -50.7 points
- GPT-6.1 Sol: Main leaderboard down 25.3 points, Code Execution -50.7 points, Material Constraints +5.8 points
Signals to Watch
- GPT-o3: Material Constraints plunged 16.7 points
- Doubao Pro: Code Execution plunged 26.2 points
- GPT-6 Astra: Main leaderboard plunged 26.3 points
- GPT-6.1 Sol: Main leaderboard plunged 25.3 points
- GPT-6 Luna: Main leaderboard plunged 20.2 points
- Grok 4: Main leaderboard plunged 33.5 points
- DeepSeek V4 Pro: Main leaderboard plunged 34.5 points
- Gemini 2.5 Pro: Main leaderboard plunged 44.8 points
- GLM-4.6: Incomplete data (missing multiple evaluation dimensions due to API failure/timeout); automatic rerun has been initiated and it is not included in the current ranking
- Qwen3 Max: Incomplete data (missing multiple evaluation dimensions due to API failure/timeout); automatic rerun has been initiated and it is not included in the current ranking
When reading this kind of Smoke brief, focus on two questions: first, whether a model has repeatedly shown the same type of weakness over multiple 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 early signals of genuine degradation, requiring follow-up review.
Data: YZ Index | Run #435 | View raw data
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接