Anthropic has launched Claude Opus 5.5, the first model in its new Claude 5.5 family. The company says the model delivers performance at the level of Claude Fable 5.1 on most work, while costing significantly less to run than the previous-generation Opus 5.
The new model is aimed heavily at coding, computer use and knowledge work. Anthropic is also highlighting improvements in writing, speed, token efficiency and safety. The company says Opus 5.5 costs 40% less to run than Opus 5 on typical workloads.
Claude Opus 5.5 Focuses On Coding And Knowledge Work
Opus 5.5 is a major step up from Opus 5, leading on agentic coding, computer use, and knowledge work. pic.twitter.com/MYI9JDAo1x
— Claude (@claudeai) September 22, 2026Anthropic says Claude Opus 5.5 leads its benchmark results in agentic coding, computer use and knowledge work. On the GDPval-AA v2.1 benchmark, which evaluates real-world work across 44 occupations, Opus 5.5 scored 1,846. Fable 5.1 scored 1,735, while Opus 5 scored 1,708.
On Terminal-Bench 4.0, Opus 5.5 achieved 66.4%, compared with 57.9% for GPT-6 Astra, according to Anthropic’s testing. However, Astra performed better on Terminal-Bench-Science 0.1, scoring 64.6% against Opus 5.5’s 58.7%.
Anthropic itself has cautioned against reading too much into small benchmark differences. The company says that at this level of capability, benchmark margins are becoming a less reliable indicator of differences in real-world performance.
Better Performance With Fewer Tokens
One of the biggest changes with Opus 5.5 is efficiency. Anthropic says the model uses fewer tokens to complete tasks while also requiring less compute to serve.
The company reports that Opus 5.5 can complete complex coding tasks with fewer steps than Opus 5. In one example, an early tester audited and fixed a 200,000-line codebase in under three hours, while Opus 5 reportedly took more than 20 hours and used 2.5 times as many tokens.
Anthropic also says Opus 5.5 generates output more than 30% faster than Opus 5. Fast mode is available through Claude Code and the Claude Platform, offering up to 2.5x the normal speed at a higher token price.
Claude Opus 5.5 Pricing
The new model is cheaper across its standard token pricing. Input tokens cost $4 per million, while output tokens cost $20 per million.
Cache reads cost just $0.20 per million tokens. Anthropic says cache reads account for a large portion of the cost in agentic and coding workloads.
For comparison, Opus 5 costs $5 per million input tokens, $25 per million output tokens and $0.50 per million cache reads. This means Opus 5.5 brings a 20% reduction in input and output token prices, while cache reads are 60% cheaper.
Claude Opus 5.5 Specifications
| Specification | Claude Opus 5.5 |
|---|---|
| Model Family | Claude 5.5 |
| Model Type | Opus |
| Positioning | Fable 5.1-level performance on most work |
| Coding | Agentic coding, codebase migrations and audits |
| Knowledge Work | Research, analysis and professional workflows |
| Computer Use | Supported |
| Input Price | $4 per 1M tokens |
| Output Price | $20 per 1M tokens |
| Cache Read | $0.20 per 1M tokens |
| Cache Write | $5 per 1M tokens |
| Cost Reduction | 40% lower typical workload cost vs Opus 5 |
| Output Speed | More than 30% faster than Opus 5 |
| Fast Mode | Up to 2.5x speed |
| Safety | Enhanced prompt-injection resistance and behavioural safeguards |
| Availability | Claude, Claude Platform, AWS, Google Cloud and Microsoft Azure |
Anthropic Also Improves Claude’s Writing
Anthropic says Opus 5.5 addresses one of the common complaints about Opus 5: overly complicated writing.
The company says the new model puts important information first, uses less jargon and follows writing instructions more closely. Early testers reportedly found its responses easier to understand during long working sessions.
This could make the difference for users who rely on Claude for reports, documentation, research and other long-form professional work.
Security And Safety Improvements
Anthropic is also putting significant focus on safety with Opus 5.5. The company says the model performed better than recent Claude models on its automated behavioural audit, which covers nearly 2,000 simulated scenarios.
Opus 5.5 also has stronger resistance to prompt injection across coding, computer use, tool use and web browsing. Anthropic says it attempted to circumvent containment boundaries around 85% less often than Opus 5 or Claude Mythos 5.1 in one internal evaluation.
The model is also launching with safeguards for sensitive biology and cybersecurity work. Verified organisations can apply for Anthropic’s relevant access programmes.
Claude Sonnet 5.5 And Haiku 5.5 Coming Soon
Claude Opus 5.5 is the first release in Anthropic’s new 5.5 family. The company says Claude Sonnet 5.5 and Claude Haiku 5.5 will follow in the coming weeks.
Opus 5.5 is already available across Anthropic’s platforms, including the Claude Platform, Amazon Web Services, Google Cloud and Microsoft Azure.
Conclusion
Claude Opus 5.5 is not simply a performance upgrade. Anthropic is putting equal emphasis on efficiency. The company says the model can deliver Fable 5.1-level performance across most work while costing substantially less to operate.
The lower token prices, reduced token usage and faster output could be particularly useful for developers running long coding and agentic workflows. For regular users, improvements in writing and higher subscription usage limits are also important changes.
Follow Us: Facebook | X | Instagram | YouTube | Pinterest
