Claude Opus 4

by Anthropic

Claude Opus 4 is positioned as the best coding model in the world, scoring 72.5% on SWE-bench and 43.2% on Terminal-bench. It sustains focus for several hours on demanding projects spanning thousands of steps, and excels at creating and maintaining memory files that keep it aware of long-term task context. Opus 4 demonstrates a 65% reduction in taking shortcuts compared to prior versions, handling multi-step problem solving that challenges competing models. It supports extended thinking with integrated tool use and parallel tool execution, enabling sophisticated agentic workflows and demanding reasoning tasks. With vision capabilities and a 200K token context window, Opus 4 is built for knowledge workers and engineers tackling high-stakes projects that require sustained reasoning and reliability across long-running work.

Key info

Input
Output
Features
Context window
200K
Max output
32K

Benchmarks

Independent benchmark scores — composite indices for reasoning, coding, and math, plus individual eval scores where available.

Intelligence Index24.3
Math Index73.3
Reasoning & knowledge
MMLU-Pro
87%
GPQA Diamond
80%
Humanity's Last Exam
12%
Coding
LiveCodeBench
64%
Agentic & tool use
Terminal-Bench Hard
31%
τ²-Bench Telecom
73%
Math & instruction following
AIME 2025
73%
IFBench
54%

Available routes

No routes currently available — Claude Opus 4 isn't routed through the Opper gateway right now. It may return.

Contact us about this model →

Available models from Anthropic

Start building with 700+ models

One API key. Every major provider. Up and running in minutes.

Get startedView Documentation