According to Artificial Analysis, the full benchmark run cost $2,486, compared with $2,304 for GPT-5.2 with maximum reasoning enabled. Opus 4.6 consumed around 58 million output tokens, roughly twice as many as Opus 4.5, but significantly fewer than the 130 million tokens used by GPT-5.2. The higher total cost reflects Anthropic’s pricing of $5 per million input tokens and $25 per million output tokens. Opus 4.6 is available via Claude.ai apps, as well as through Anthropic’s API, Google Vertex, AWS Bedrock, and Microsoft Azure.
Claude Opus 4.6 is currently the best AI model—at least until Artificial Analysis benchmarks OpenAI’s Codex 5.3, which is expected to take the lead in coding tasks. For now, Anthropic’s new model ranks first in the Artificial Analysis Intelligence Index, which aggregates results from ten benchmarks covering coding, agent-based tasks, and scientific reasoning. Opus 4.6 leads in three categories: agentic work tasks, terminal-based coding, and physics research problems
ES 
EN