Claude Opus 4.6
Anthropic🇺🇸 United States
Anthropic's most intelligent model. Leads on frontier reasoning, coding, and research.
Context window200K tokens
Input / 1M tokens$15
Output / 1M tokens$75
Version History
claude-opus-4-6-20260217
Claude Opus 4.6 — major GPQA and reasoning improvements; ARC-AGI-2 jump from 37.6% to 68.8%.
Benchmark Scores
Full leaderboard →89.0%
AIME 2024
91.0%
AIME 2025
95.4%
DocVQA
91.3%
GPQA
92.0%
HumanEval
71.2%
LiveCodeBench
94.2%
MATH
90.2%
MMLU
85.1%
MMLU-Pro
82.0%
MMMU
38.0 tokens_per_sec
Speed (tok/s)
80.8%
SWE-bench Verified