Claude Sonnet 4.6 Leads with 89.8: 2026-10-24 Smoke Quick-Test Data Brief

The 2026-10-24 YZ Index Smoke quick test covered 13 models, with Claude Sonnet 4.6 ranking first for the day at 89.8 points. Smoke is a daily 10-question quick test suited to observing short-term signals; it is not equivalent to the conclusions of the weekly Full 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 treated as monitoring signals rather than as long-term verdicts on model capability.

Daily Ranking

Data Interpretation

The top two on today's main leaderboard are both Claude-series models. Claude Sonnet 4.6 scored 97 on code execution and 81 on material constraints for a main-leaderboard score of 89.8, while Claude Opus 4.7 matched it with 97 on code execution paired with 72.6 on material constraints for 86.02. Grok 4 and Doubao Pro followed close behind: the former posted 74.2 on code execution and 81 on material constraints for 77.26 on the main leaderboard, while the latter posted 72 on code execution and 83.3 on material constraints for 77.09. The leading models differ in how they balance strength across code execution and material constraints: the Claude series holds a high position on code execution, while Doubao Pro is slightly stronger on material constraints.

GPT-6.1 Sol fell 33.3 points on the main leaderboard, with code execution down 53 points and material constraints down 9.1 points; Gemini 3.1 Pro fell 28.8 points on the main leaderboard with code execution down 53 points; GPT-6 Astra fell 27.4 points on the main leaderboard with code execution down 47.8 points; DeepSeek V4 Pro fell 25 points on the main leaderboard, with code execution down 53 points and material constraints up 9.3 points. Claude Sonnet 4.6, by contrast, rose 11.5 points on the main leaderboard with material constraints up 29.3 points. These single-day movements may stem from fluctuations in question sampling or may reflect genuine changes in model performance; both require follow-up runs using the same methodology for confirmation.

The Smoke quick test is a small-sample, single-day signal. The score structure and changes above reflect only the data for that day, and no long-term judgment is being made for now.

Key Changes

• GPT-6.1 Sol: main leaderboard down 33.3 points, code execution −53 points, material constraints −9.1 points
• Gemini 3.1 Pro: main leaderboard down 28.8 points, code execution −53 points
• GPT-6 Astra: main leaderboard down 27.4 points, code execution −47.8 points
• DeepSeek V4 Pro: main leaderboard down 25 points, code execution −53 points, material constraints +9.3 points
• Claude Sonnet 4.6: main leaderboard up 11.5 points, material constraints +29.3 points

Signals to Watch

• No publishable anomaly signals were retained for this test.

When reading this kind of Smoke brief, the focus should be on two questions: first, whether a given model exposes the same type of weakness across multiple consecutive days; second, whether its integrity rating moves from pass into warn or fail. Large single-day changes in execution or constraint scores may come from question sampling, or they may be early signals of genuine degradation, and they require follow-up runs for confirmation.

Data source: YZ Index | Run #393 | View raw data