GPT-6.1 Sol Leads with 86.05: 2026-10-26 Smoke Quick Test Data Brief

On 2026-10-26, the YZ Index Smoke quick test covered 13 models, with GPT-6.1 Sol taking first place for the day at 86.05 points. Smoke is a daily 10-question quick test, suited to observing short-term signals and 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 used as monitoring signals rather than as long-term conclusions about model capability.

Rankings for the Day

Data Interpretation

Today's top three models on the main leaderboard show a clear divergence in how they pair code execution with material constraints. GPT-6.1 Sol took 86.05 on the main leaderboard with a code execution score of 100 and material constraints of 69; GPT-6 Astra also scored 100 on code execution but only 63.3 on material constraints, for 83.49 on the main leaderboard; Claude Sonnet 4.6 scored 72.4 on code execution and 79.8 on material constraints, for 75.73 on the main leaderboard. This combination of strong code execution with moderate material constraints gives the top two a clear advantage under the main leaderboard formula, while Claude Sonnet 4.6, which is relatively stronger on material constraints, ranks third overall.

Several models fell sharply compared with the previous run under the same methodology. GPT-5.5 dropped 63 points on the main leaderboard, 92.6 points on code execution, and 26.9 points on material constraints; Doubao Pro dropped 45.6 points on the main leaderboard, 68.6 points on code execution, and 17.4 points on material constraints; GPT-o3 dropped 41.3 points on the main leaderboard and 70.5 points on code execution; Gemini 3.1 Pro and Grok 4 also fell 38.4 and 38.1 points on the main leaderboard, respectively. These changes may stem from fluctuations in question sampling, or may reflect genuine degradation in model performance; because the Smoke test is a small-sample, single-day signal, subsequent runs are needed to confirm stability.

The remaining models, such as DeepSeek V4 Pro with 0 on code execution and 93.3 on material constraints for 41.99 on the main leaderboard, and GPT-o3 with 0 on code execution and 78.8 on material constraints for 35.46, all display a structure dominated by material constraints. Overall, the differences in score structure between the leading models and the models that moved sharply are mainly concentrated in the balance between code execution and material constraints; the current data is for reference for the day only.

Key Changes

• GPT-5.5: main leaderboard down 63 points, code execution -92.6 points, material constraints -26.9 points
• Doubao Pro: main leaderboard down 45.6 points, code execution -68.6 points, material constraints -17.4 points
• GPT-o3: main leaderboard down 41.3 points, code execution -70.5 points, material constraints -5.7 points
• Gemini 3.1 Pro: main leaderboard down 38.4 points, code execution -50 points, material constraints -24.1 points
• Grok 4: main leaderboard down 38.1 points, code execution -52.7 points, material constraints -20.2 points

Signals to Watch

• No publishable anomaly signals were retained in this run.

When reading this kind of Smoke brief, the focus should be on two questions: first, whether a given model repeatedly exposes the same type of weakness over consecutive 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 genuine degradation, and require follow-up runs to confirm.

Data source: YZ Index | Run #396 | View raw data