Ordalin
Submit a tool

Reviewed tool profile

Claude Opus 5.5

Anthropic’s Opus-tier Claude model for long-running agentic coding and knowledge work, at $4/$20 per million tokens.

Sources checked

Visit website ↗
Pricing
API usage is billed per million tokens: $4 input and $20 output, with cache reads at $0.20, 5-minute cache writes at $5 and 1-hour cache writes at $8. The Batch API takes 50% off input and output. Fast mode costs $8 input and $40 output, and US-only inference costs 1.1x. Official source 1 ↗ · Official source 2 ↗ · Official source 3 ↗
Available as
Available on the Claude Platform API as claude-opus-5-5, and on Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Platform on AWS. · Available in the Claude apps for Pro, Max, Team and Enterprise users, and in Claude Code. Official source 1 ↗ · Official source 2 ↗ Official source 1 ↗
Primary group
Coding & Development

What it does

Claude Opus 5.5 is Anthropic’s Opus-tier model for long-running agentic coding and knowledge work, released on September 22, 2026 as the first model of the Claude 5.5 family. Anthropic says it performs at the level of Claude Fable 5.1 on most work, leads its benchmarks in agentic coding, computer use and knowledge work, and costs about 40% less to run than Opus 5 at default settings.

Official source 1 ↗ · Official source 2 ↗

Features

  • Long-running agentic coding. Built for long, sprawling jobs such as codebase-wide migrations and audits. Anthropic reports 66.4% on Terminal-Bench 4.0 and 54.6% on FrontierCode at default effort. Official source 1 ↗
  • 1M context and 128K output. A 1M-token context window and 128K output tokens, or up to 300K output on the Batch API in beta. Reliable knowledge cutoff June 2026. Official source 1 ↗
  • Effort control. Adaptive thinking is always on; the effort parameter sets thinking depth from low to max, with medium as the default. Official source 1 ↗ · Official source 2 ↗
  • Fast mode. Up to 2.5x faster output for $8 input and $40 output per million tokens, in research preview on the Claude API only. Official source 1 ↗ · Official source 2 ↗
  • Knowledge work and research. Scores 1846 Elo on GDPval-AA v2.1; in Anthropic’s research test 16 of 18 reports passed a no-invented-figures bar. Official source 1 ↗
  • Computer use. Anthropic reports 81.8% on OSWorld 2.1 (partial). Official source 1 ↗
  • Clearer writing. Puts the most important information first, avoids jargon and follows writing rules you give it. Official source 1 ↗

Pricing & free limits

API — standard

$4 input / $20 outputUSD per million tokens

Cache reads $0.20, 5-minute cache writes $5 and 1-hour cache writes $8 per million tokens. Batch API 50% off. Official source 1 ↗ · Official source 2 ↗

Claude Pro, Max, Team and Enterprise

Included in planSubscription

Opus 5.5 is available in the Claude apps on these paid plans. Official source 1 ↗

Billing details

Anthropic says Opus 5.5 costs about 40% less than Opus 5 on typical workloads at default settings, from lower per-token prices and fewer tokens per task. Official source 1 ↗

US-only inference costs 1.1x the standard input and output price. Official source 1 ↗

Anthropic raised five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans with this release. Official source 1 ↗

Free access & limits

Free access and its limits were not established by the checked official sources.

Best for & limitations

Run long, unattended codebase migrations, audits and refactors. Official source 1 ↗

Write researched reports and financial analysis where invented figures are unacceptable. Official source 1 ↗

Lead an agent that delegates to cheaper subagents, at medium effort by default. Official source 1 ↗

Automate computer-use and browser tasks. Official source 1 ↗

  • Thinking cannot be turned off. Adaptive thinking is always on, and four breaking changes affect code written for Opus 5. Official source 1 ↗
  • Cybersecurity and biology safeguards. Most cybersecurity tasks are re-routed to Opus 4.8, and biology work uses Fable 5.1’s safeguards unless an organisation is verified. Official source 1 ↗
  • Fast mode is limited. Fast mode is a research preview on the Claude API only, not on Bedrock, Google Cloud or Microsoft Foundry. Official source 1 ↗
  • Preserved thinking. API accounts created on or after August 31, 2026 cannot edit prior context to extract the model’s reasoning. Official source 1 ↗
  • Benchmarks at max effort. Anthropic’s headline scores use max effort unless noted, and it says the model often suspects it is being evaluated. Official source 1 ↗
Claude Opus 5.5 product screenshot
Website previewFixed desktop viewport · 1440 × 900

Editorial

Shortlists, alternatives and comparisons that include Claude Opus 5.5.

ComparisonArgon vs Opus: Gemini 4 Argon or Claude Opus 5.5 for coding agents and long tasks?Gemini 4 Argon and Claude Opus 5.5 compared on API price before and after December 1, context, output length, thinking control, platforms and limits.

Similar tools

Published tools that share its categories.