Claude Opus 4 (Thinking): benchmark results
Claude Opus 4 evaluated with thinking enabled. Provider: Anthropic. Released 2025-05-22. Access: API.
Aggregate Rating 1681 ± 13, rank #417 of 2168 rated models, from 89 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SuperCLUE General (July 2025) - Hallucination Control | 90.24 | Score | 100 |
| BenchTable | 82.8 | Total Score (%) | 99.1 |
| BenchTable - Reasoning | 91.4 | Weighted Score (%) | 99.1 |
| AGI-Eval Community - Cognition (English) | 89.88 | Accuracy (%) | 98.9 |
| BenchTable - STEM | 87.2 | Weighted Score (%) | 98.3 |
| BenchTable - Tech | 89.7 | Weighted Score (%) | 97.5 |
| SuperCLUE General (May 2025) - Text Understanding and Creation | 80.62 | Score | 97.4 |
| AA MMLU-Pro | 87.32 | Accuracy (%) | 97.1 |
| Ducky Bench (Saxo Frog) | 1280 | ELO | 95.8 |
| SuperCLUE General (May 2025) - Agent | 70.61 | Score | 94.9 |
| SuperCLUE General (May 2025) - Overall | 65.74 | Score | 94.9 |
| SuperCLUE General (May 2025) - Precise Instruction Following | 59.1 | Score | 94.9 |
Interactive version: theaggregate.ai/model?slug=claude-opus-4-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.