Claude Opus 4.6
FrontierMultimodalAPI Available
Benchmark Scores
94.5/100MMLU94.8
HumanEval95.5
Math91.8
Reasoning95
Coding95.2
About Claude Opus 4.6
Claude Opus 4.6 is Anthropic's most powerful model, featuring enhanced extended thinking and a 400K context window. It excels at complex coding, research synthesis, and multi-step agentic tasks with industry-leading safety.
Strengths
- +Exceptional extended thinking
- +Best-in-class coding
- +Superior instruction following
- +400K context with perfect recall
Weaknesses
- -Premium pricing
- -Higher latency with extended thinking
- -Occasionally overly cautious on edge cases
Pricing
Per 1M tokens
Input$15.00
Output$75.00
Quick Facts
- Context Window
- 400K
- Parameters
- Undisclosed
- Release Date
- 2026-03-01
- Category
- Frontier
Compare Models
See how Claude Opus 4.6 stacks up against other models