Claude Opus 5.5
Anthropic · September 22, 2026
● activeCloseddecoder onlymultimodal
Context Window1M tokens
Description
Anthropic's premier frontier milestone featuring proactive credit assignment, formal tool verification, and deep mathematical reasoning.
Benchmark Scores
MMLUMassive Multitask Language Understanding — 57 subjects
94.1%HumanEvalCode generation pass@1 — Python problems
96.7%MATHMATH benchmark — competition-level problems
98.6%GPQAGraduate-level science QA
83.2%SWE-benchReal-world software engineering
75.1%Key Innovations
Reasoning
ReasoningStructured step-by-step problem solving, often using chain-of-thought or tree-of-thought approaches.
Agentic
AgenticModels that can autonomously plan, execute multi-step tasks, use tools, and self-correct without human intervention.
Constitutional AI
Constitutional AIAnthropic's approach to AI safety where the model critiques and revises its own outputs against a set of principles.
Tool Use
Tool UseAbility to call external tools, APIs, and functions — enabling web browsing, code execution, and real-world actions.
Enabled By
NVIDIA Vera Rubin (NV-VR200)NVIDIA · September 2026
50 PFLOPS FP4 Tensor Core