Grok 4 Leads with 83.51: 2026-11-22 Smoke Quick Test Data Brief

The 2026-11-22 YZ Index Smoke Quick Test covered 13 models, and Grok 4 ranked first for the day with 83.51 points. Smoke is a daily 10-question quick test, suitable for observing short-term signals; it is not equivalent to conclusions from the Full weekly 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, a single-day score is better used as a monitoring signal rather than as a long-term verdict on model capability.

Daily Ranking

RankModelMain LeaderboardCode ExecutionMaterial ConstraintsIntegrity
#1Grok 483.5188.577.4pass
#2GPT-o378.471.686.7pass
#3Claude Opus 4.778.0999.152.4pass
#4Doubao Pro73.1773.872.4pass
#5GPT-6.1 Sol70.8874.366.7pass
#6GPT-6 Luna67.1371.461.9pass
#7GPT-6 Astra66.7974.357.6pass
#8GPT-6 Sol64.6374.352.8pass
#9DeepSeek V4 Pro62.335077.4pass
#10Claude Sonnet 4.661.5648.677.4pass
#11Gemini 2.5 Pro57.525066.7pass
#12Gemini 3.1 Pro57.525066.7pass
#13GPT-5.554.1649.360.1pass

Data Interpretation

Today's top three models on the main leaderboard show different profiles in the balance between Code Execution and Material Constraints. Grok 4 scored 83.51 on the main leaderboard with Code Execution 88.5 and Material Constraints 77.4, showing relatively balanced capabilities; GPT-o3 scored 78.4 on the main leaderboard with Code Execution 71.6 and Material Constraints 86.7, with Material Constraints relatively stronger; Claude Opus 4.7 scored 78.09 on the main leaderboard with Code Execution 99.1 and Material Constraints 52.4, with standout Code Execution but relatively low Material Constraints. Doubao Pro scored 73.17 on the main leaderboard, with Code Execution 73.8 and Material Constraints 72.4, a relatively balanced combination.

Several models showed notable score movements. GPT-o3's Code Execution plunged by 27.2 points; Claude Opus 4.7's Material Constraints plunged by 17.6 points; Doubao Pro's main leaderboard score plunged by 17.9 points, with Code Execution down 25 points; GPT-6 Luna's main leaderboard score plunged by 13 points, with Code Execution down 27.9 points; DeepSeek V4 Pro's Code Execution plunged by 25 points; Claude Sonnet 4.6's Code Execution plunged by 25.2 points; Gemini 2.5 Pro's main leaderboard score plunged by 12.2 points, with Code Execution down 25 points; and Gemini 3.1 Pro's main leaderboard score plunged by 15.2 points. These changes may stem from question sampling fluctuations or may reflect real degradation, and follow-up runs are needed for confirmation. GLM-4.6 and Qwen3 Max had incomplete data due to API failures and were not included in this period's ranking.

The Smoke quick test is a small-sample, single-day signal; the observations above reflect only that day's data performance and do not support long-term inferences.

Key Changes

  • Doubao Pro: Main Leaderboard down 17.9 points, Code Execution down 25 points, Material Constraints down 9.3 points
  • Gemini 3.1 Pro: Main Leaderboard down 15.2 points, Code Execution down 25 points
  • GPT-6 Luna: Main Leaderboard down 13 points, Code Execution down 27.9 points, Material Constraints up 5.2 points
  • Gemini 2.5 Pro: Main Leaderboard down 12.2 points, Code Execution down 25 points
  • Claude Sonnet 4.6: Main Leaderboard down 7.5 points, Code Execution down 25.2 points, Material Constraints up 14.1 points

Signals to Watch

  • GPT-o3: Code Execution plunged by 27.2 points
  • Claude Opus 4.7: Material Constraints plunged by 17.6 points
  • Doubao Pro: Main Leaderboard plunged by 17.9 points
  • GPT-6 Luna: Main Leaderboard plunged by 13 points
  • DeepSeek V4 Pro: Code Execution plunged by 25 points
  • Claude Sonnet 4.6: Code Execution plunged by 25.2 points
  • Gemini 2.5 Pro: Main Leaderboard plunged by 12.2 points
  • Gemini 3.1 Pro: Main Leaderboard plunged by 15.2 points
  • GLM-4.6: Data incomplete (missing integrity and communication dimensions; API failure/timeout); automatically queued for rerun and not included in this period's ranking
  • Qwen3 Max: Data incomplete (missing multiple evaluation dimensions; API failure/timeout); automatically queued for rerun and not included in this period's ranking

When reading this type of Smoke brief, focus on two questions: first, whether a model has shown the same type of weakness for multiple 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 early signals of real degradation, requiring follow-up runs for confirmation.


Data source: YZ Index | Run #438 | View raw data

This article is from Winzheng Index blog, translated in full by Winzheng (winzheng.com). Click here to view the original When republishing the translation, please credit the source. Thank you!