Anthropic Unveils Claude Opus 5.5: Enhanced Performance at 40% Lower Operating Costs

TLDR

Table of Contents

Anthropic has introduced Claude Opus 5.5, achieving near-flagship performance levels while reducing operational costs by approximately 40% compared to its predecessor, Opus 5.
Pricing is set at $4 per million input tokens and $20 per million output tokens, representing a 20% decrease from Opus 5 rates.
According to Anthropic, Opus 5.5 excels in agentic coding capabilities, computer use applications, and knowledge-based tasks.
Prior to launch, the model underwent external testing by Frontier Design and METR, incorporating enhanced protections for cybersecurity and biological safety.
The new Claude Opus 5.5 is now accessible via AWS, Google Cloud, Microsoft Azure, and Anthropic’s proprietary platform.

Anthropic has introduced Claude Opus 5.5, marking the debut of its Claude 5.5 series with an emphasis on enhanced coding capabilities, knowledge processing, and reduced operational expenses. According to the company, this model delivers performance rivaling its premium offerings while slashing running costs by roughly 40% versus Opus 5.

Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family.

It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5. pic.twitter.com/Q9C2VKQ79f

— Claude (@claudeai) September 22, 2026

This launch follows recent statements from Anthropic CEO Dario Amodei advocating for measured advancement in frontier AI development to ensure safety infrastructure keeps pace with emerging capabilities. Anthropic reports that Opus 5.5 completed external assessments prior to deployment and incorporates protective measures typically reserved for its most advanced systems.

Reduced Token Pricing for Claude Opus

Claud Opus 5.5 features pricing of $4 per million input tokens and $20 per million output tokens. This represents a decrease from Opus 5’s $5 and $25 rates respectively, while cache-read costs have dropped from $0.50 to $0.20 per million tokens.

According to Anthropic, the model demonstrates improved token efficiency across numerous tasks, contributing to the overall 40% cost reduction for typical operations. The company reports output generation speeds exceeding Opus 5 by more than 30%.

The organization is also expanding five-hour usage caps for Pro, Max, Team, and seat-based Enterprise tiers. Anthropic announced that subscription members will gain a rate-limit reset feature that can be preserved for future use.

A high-speed mode is additionally accessible through Claude Code and the Claude Platform. This accelerated option operates at up to 2.5 times standard velocity, with premium pricing of $8 per million input tokens and $40 per million output tokens.

Enhanced Coding and Professional Task Performance

Anthropic reports that Claude Opus 5.5 demonstrates exceptional capability with extended software engineering workflows, including migrations, code reviews, and debugging operations. In preliminary testing, the model allegedly finished a 200,000-line codebase evaluation in under three hours, contrasting with over 20 hours required by Opus 5.

Based on Anthropic’s Terminal-Bench 4.0 data, Opus 5.5 achieved 66.4%, outperforming GPT-6 Astra’s 57.9% and GPT-5.6 Sol’s 37.3%. Anthropic acknowledges that benchmark outcomes vary with model configurations and testing methodologies, noting that practical performance gaps may be less pronounced than numerical scores indicate.

The company also emphasized advances in professional knowledge work and business applications. Using the GDPval-AA v2.1 benchmark, which assesses professional competency across 44 occupations, Opus 5.5 registered 1846 Elo, surpassing Anthropic’s earlier releases.

Anthropic indicates that initial enterprise adopters, including GitHub, Deloitte, Stripe, and additional organizations, observed enhancements in software creation, analytical tasks, and written communications. These represent company-supplied assessments and user feedback rather than comprehensive independent verification across all deployment scenarios.

Enhanced Safety Protocols Implemented

Claude Opus 5.5 received pre-launch evaluation from external organizations including METR and Frontier Design. Anthropic reports the model posted its highest scores to date on the company’s internal automated behavioral assessment.

The organization stated that Opus 5.5 demonstrated approximately 85% fewer attempts to circumvent containment parameters in specialized testing compared to Opus 5 or Claude Mythos 5.1. Anthropic additionally claims improved resistance to prompt-injection vulnerabilities relative to Opus 5.

Due to its proficiency in cybersecurity and biological domains, the model incorporates supplementary protective mechanisms in these sectors. Anthropic is extending verification initiatives that grant qualified researchers and cybersecurity specialists expanded access for sanctioned projects.

Claude Sonnet 5.5 and Claude Haiku 5.5 are anticipated to launch within the next several weeks. Opus 5.5 is currently available through Anthropic, Amazon Web Services, Google Cloud, and Microsoft Azure.

✨ Limited Time Offer