GPT-5.5 and GPT-6 Sol Tie at 92.44 Points: 2026-11-16 Smoke Quick Test Data Brief

The 2026-11-16 YZ Index Smoke quick test covered 14 models, with GPT-5.5 and GPT-6 Sol tied for first place that day at 92.44 points. Smoke is a daily 10-question quick test, suitable for observing short-term signals and 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, single-day scores are better treated as monitoring signals rather than long-term conclusions about model capability.

Daily Rankings

Data Interpretation

Today's YZ Index Smoke quick test shows that GPT-5.5 and GPT-6 Sol tied at 92.44 points on the main leaderboard, with both scoring 95 in code execution and 89.3 in material constraints, reflecting a strong combination of code execution and material constraints. GPT-6 Luna scored 88.75 on the main leaderboard, with 88.3 in code execution and 89.3 in material constraints, also relying on material constraints to maintain a relatively high position. Claude Opus 4.7 scored 82.43 on the main leaderboard, with 76.8 in code execution and 89.3 in material constraints; Grok 4 scored 81.99 on the main leaderboard, with 76 in code execution and 89.3 in material constraints. The models after these generally show a pattern of lower code execution while material constraints remain at 89.3.

GPT-o3 fell 25.9 points on the main leaderboard and 47 points in code execution; Doubao Pro fell 24.6 points on the main leaderboard, 24.2 points in code execution, and 25 points in material constraints; Claude Sonnet 4.6 fell 24.5 points on the main leaderboard, 24 points in code execution, and 25 points in material constraints. These anomalous signals may come from sampling variation in questions or may be signs of genuine degradation, and require follow-up runs to recheck. GPT-6 Astra's code execution plunged 27.3 points, GPT-6.1 Sol's code execution plunged 27.3 points, Gemini 3.1 Pro's material constraints plunged 25 points, and Gemini 2.5 Pro's main leaderboard score fell 8.9 points; these also require multiple rounds of rechecking to distinguish fluctuation from degradation.

As a small-sample single-day signal, Smoke should be interpreted with restraint; no conclusions are drawn about long-term performance.

Key Changes

• GPT-o3: Main leaderboard down 25.9 points; code execution -47 points
• Doubao Pro: Main leaderboard down 24.6 points; code execution -24.2 points; material constraints -25 points
• Claude Sonnet 4.6: Main leaderboard down 24.5 points; code execution -24 points; material constraints -25 points
• Gemini 2.5 Pro: Main leaderboard down 8.9 points; material constraints -25 points
• GPT-6 Luna: Main leaderboard down 6.4 points; code execution -11.7 points

Signals to Watch

• GPT-6 Astra: Code execution plunged -27.3 points
• GPT-6.1 Sol: Code execution plunged -27.3 points
• Claude Sonnet 4.6: Main leaderboard plunged -24.5 points
• GPT-o3: Main leaderboard plunged -25.9 points
• Gemini 3.1 Pro: Material constraints plunged -25 points
• Gemini 2.5 Pro: Main leaderboard plunged -8.9 points
• Doubao Pro: Main leaderboard plunged -24.6 points
• Qwen3 Max: Data incomplete (missing several evaluation dimensions due to API failure/timeout); an automatic rerun has been initiated, and it is not included in this period's rankings

When reading this kind of Smoke briefing, focus on two questions: first, whether a model has repeatedly shown the same type of weakness over 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 genuine degradation, requiring follow-up runs to recheck.

Data source: YZ Index | Run #429 | View raw data