Claude Opus 4.7 Leads at 95.19 Points: 2026-08-03 Smoke Quick Test Data Briefing

On 2026-08-03, the YZ Index Smoke quick test covered 11 models, with Claude Opus 4.7 ranking first at 95.19 points. Smoke is a daily 10-question quick test suitable for observing short-term signals, and is not equivalent to the Full weekly ranking conclusions.

This Smoke evaluation only covers two main ranking dimensions: code execution and material constraints. The main ranking formula is 0.55 × code execution + 0.45 × material constraints. Due to the small daily sample size, single-day scores are better used as monitoring signals rather than definitive long-term judgments of model capabilities.

Daily Ranking

Data Interpretation

In today's Smoke quick test, Claude Opus 4.7 ranked first with a main score of 95.19, forming an advantage through its combination of 100 in code execution and 89.3 in material constraints. Grok 4 followed closely with a main score of 89.31, code execution of 97.9, and material constraints of 78.8. Gemini 2.5 Pro scored 98.7 in code execution but only 53.8 in material constraints, yielding a main score of 78.5. GLM-4.6 presented a combination of 89.3 in material constraints and 58.3 in code execution, with a main score of 72.25, illustrating different weight balances among top models under the 0.55 × code execution + 0.45 × material constraints formula.

Multiple models showed notable declines. Doubao Pro's main score dropped 40.8 points, with code execution down 36.4 points and material constraints down 46.2 points. Qwen3 Max's main score fell 37.2 points, with code execution down 22 points and material constraints down 55.7 points. Gemini 3.1 Pro's main score fell 33.8 points, with code execution down 36.4 points and material constraints down 30.7 points. These changes may stem from single-day small-sample sampling fluctuations, or may reflect genuine performance degradation, requiring subsequent runs under the same methodology for confirmation.

Claude Sonnet 4.6 and DeepSeek V4 Pro also recorded main score declines of 31.5 and 24.3 points respectively, with both code execution and material constraints trending downward, while overall structure maintained a pass integrity record. As Smoke is a small-sample single-day signal, the current data is for reference only; no long-term inferences should be made.

Key Changes

• Doubao Pro: Main score down 40.8 points, code execution -36.4 points, material constraints -46.2 points
• Qwen3 Max: Main score down 37.2 points, code execution -22 points, material constraints -55.7 points
• Gemini 3.1 Pro: Main score down 33.8 points, code execution -36.4 points, material constraints -30.7 points
• Claude Sonnet 4.6: Main score down 31.5 points, code execution -44 points, material constraints -16.2 points
• DeepSeek V4 Pro: Main score down 24.3 points, code execution -19 points, material constraints -30.7 points, integrity warn→pass

Signals to Watch

• No release-worthy anomaly signals were retained this time.

When reading this type of Smoke briefing, the focus should be on two questions: first, whether a particular model has been exposed to the same type of weakness for multiple consecutive days; second, whether the integrity rating has moved from pass to warn or fail. Large single-day swings in execution or constraint scores may come from question sampling, or may be early signals of genuine degradation, requiring subsequent run reviews for confirmation.

Data source: YZ Index | Run #258 | View raw data