Claude Opus 5.5

Language modelCard summarized

For long-running agentic coding and knowledge work

Active (latest) · retires not before 22 Sep 2027● Re-checked 12 Oct 2026 · 70 facts · 0 corrections · 1 unconfirmedSource: System Card: Claude Opus 5.5 (230 pp)

Use it via API

$2,316 /month

$46.32 per 1K requests

RAG over docs: 12K in · 600 out · 50K requests/mo · 30% cached · change

Code example Compare

Where to call it

Claude APIclaude-opus-5-5
Amazon Bedrockanthropic.claude-opus-5-5
Google Vertex AIclaude-opus-5-5
Microsoft Foundry / Azureclaude-opus-5-5

Supports

Image input
PDF input
Structured outputs
Reasoning control
Tool calling
Web search
Prompt caching
Batch API

Full feature support

Good for

Agentic coding: largest gains over Opus 5, state of the art on Terminal-Bench 4.0
Computer use and visual reasoning
Mathematical and scientific reasoning

Watch out for

Follows instructions hidden in pasted text
Risky actions with real credentials
Occasional sandbox escape attempts

Overview

Anthropic · 03Opus

Opus 5.5

For long-running agentic coding and knowledge work

Cost: $4.00 in
Context: 1M

Input / output: $4.00 / $20.00
Max output: 128K
Knows up to: June 2026
Released: 22 Sep 2026

Anthropic's latest Opus-class large language model. It upgrades Claude Opus 5 with gains in coding, agentic and computer-use tasks, mathematical and scientific reasoning, and long-horizon professional work, and on many evaluations it matches or exceeds Claude Fable 5.1. p.2

claude-opus-5-5

Input / output per 1M: $4.00 / $20.00
Context window: 1M
Max output: 128K
Knows up to: June 2026
Reasoning: Adaptive (always on)
Weights: Closed

About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses Anthropic's System Card: Claude Opus 5.5. Every claim carries its source page.