# Expert

9 models

World-leading HLE score of 52.15%

Maximum rigor for high-stakes, mission-critical work. Uses 9 models with up to extra-high thinking effort for research-grade analysis, exhaustive verification, and professional-standard deliverables in regulated domains.

## Technical Specifications

### Model Count

9 models

### Max Thinking Effort

Extra High

### Max Thinking Budget

4,096 tokens

### Synthesizer

Claude Opus 4.7

## How It Works

### Multi-Model Synthesis

This mode uses 9 AI models working in parallel to analyze your request. Their responses are then synthesized by [Claude Opus 4.7](/content/models/anthropic-claude-opus-4.7) to produce the final answer. This approach improves accuracy by leveraging multiple perspectives and cross-validating conclusions.