Claude Opus 4.1: benchmark results
Anthropic Opus-tier Claude model, an incremental Opus 4 upgrade focused on coding and agentic tasks. Provider: Anthropic. Released 2025-08-05. Access: API.
Aggregate Rating 1733 ± 6, rank #140 of 1656 rated models, from 459 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| FutureSearch DRB - Find Dataset | 0.71 | Average Score | 100 |
| ImplicitMemBench - Procedural Memory | 76.67 | First-attempt accuracy (%) on the 100 procedural-memory item | 100 |
| LLM Stats (TAU-bench Retail; Claude extended thinking; prompt addendum; 100 steps) | 82.4 | Score (%) | 100 |
| MobiFlow - App Interference Completion | 70.1 | Completion rate (%) under app interference (pop-ups and othe | 100 |
| Nejumi 4 - ALT - Truthfulness | 88.5 | Score (%) | 100 |
| Nejumi 4 - JTruthfulQA | 78.01 | Score (%) | 100 |
| PCB-Bench - Placement Macro CQ | 93.3 | Accuracy (%) | 100 |
| PCB-Bench - Placement Micro CQ | 94.35 | Accuracy (%) | 100 |
| PCB-Bench - Routing Micro CQ | 92.32 | Accuracy (%) | 100 |
| PCB-Bench - Routing Micro QA SBERT | 57.38 | SBERT similarity (%) | 100 |
| Sand-Glass (Prompted Compression) - Chinese-English Translation Density | 0.38 | BLEU-rho (back-translation BLEU divided by the target/source | 100 |
| Sand-Glass (Prompted Compression) - Chinese-German Translation Density | 0.29 | BLEU-rho (back-translation BLEU divided by the target/source | 100 |
Interactive version: theaggregate.ai/model?slug=claude-opus-4-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.