Anthropic has launched Claude Sonnet 5.5, the second model in its Claude 5.5 family, with a focus on faster responses, improved coding and knowledge-work performance, and lower costs per completed task. Announced September 28, 2026, the new model keeps the same base token pricing as Sonnet 5 while generating output more than 30% faster and, according to Anthropic’s testing, costing up to 30% less per task.
Sonnet 5.5 is positioned below Claude Opus 5.5 in Anthropic’s model lineup. The company describes it as particularly suited to well-defined everyday tasks, software development, document creation, presentations and spreadsheets, while Opus 5.5 remains aimed at more complex work requiring sustained judgment.
Quick Summary
- Anthropic launched Claude Sonnet 5.5 on September 28, 2026.
- It is the second model in the Claude 5.5 family, alongside Claude Opus 5.5.
- Anthropic says Sonnet 5.5 generates output more than 30% faster than Sonnet 5.
- The model keeps the same headline API token pricing as Sonnet 5.
- Anthropic reports up to 30% lower cost per task because Sonnet 5.5 typically uses fewer tokens.
- Coding performance improved substantially on Anthropic’s published benchmarks.
- Sonnet 5.5 is also designed for documents, presentations, spreadsheets and UI-related work.
- It introduces stronger cybersecurity safeguards within the Sonnet line.
- Sonnet 5.5 is available across Claude and major developer platforms.
- Claude Haiku 5.5 is expected to join the family in the coming weeks.
What Is New in Claude Sonnet 5.5?
The most significant change is not a new token price. Instead, Anthropic is emphasizing the amount of work Sonnet 5.5 can complete with fewer tokens and less processing time.
Anthropic says Sonnet 5.5 generates outputs 30% or more faster than Sonnet 5, making it the fastest Sonnet model the company has released. At the same time, its input and output token prices remain at $2 and $10 per million tokens respectively. Cache reads cost $0.20 per million tokens.
The lower effective cost comes from improved efficiency. Anthropic says Sonnet 5.5 typically requires fewer tokens to complete the same work, resulting in up to 30% lower cost per task in its testing.
| Metric | Claude Sonnet 5.5 | Claude Sonnet 5 |
|---|---|---|
| Input price / 1M tokens | $2 | $2 |
| Output price / 1M tokens | $10 | $10 |
| Cache reads / 1M tokens | $0.20 | $0.20 |
| Output speed | 30%+ faster | Baseline |
| Terminal-Bench 4.0 | 70.6% | 10.3% |
| CursorBench 4.0 | 55.5% | 34.1% |
| OSWorld 2.1 | 80.1% partial | 57.0% partial |
The benchmark figures are Anthropic’s published results and should be interpreted within the specific evaluation methodology and effort settings used. Anthropic also notes that benchmark performance represents only one aspect of model capability.
Claude Sonnet 5.5 Targets Coding and Everyday AI Work
Coding is one of the areas where Anthropic reports some of the largest improvements.
On Terminal-Bench 4.0, an evaluation of agentic coding tasks performed through a command-line interface, Sonnet 5.5 scored 70.6%, compared with 10.3% for Sonnet 5. On CursorBench 4.0, which uses ambiguous, multi-file coding tasks derived from real Cursor sessions, Sonnet 5.5 scored 55.5%, compared with 34.1% for its predecessor.
Anthropic also says Sonnet 5.5 can batch tool calls more effectively, reducing the number of steps needed to complete some coding tasks. This can matter for developers using AI coding agents, where both response time and token consumption contribute to the overall cost of a workflow.
At particular effort levels, Anthropic reports that Sonnet 5.5 can exceed Sonnet 5’s best benchmark scores at roughly a tenth of the cost per task. These figures vary by benchmark and effort setting rather than representing a universal tenfold improvement in cost efficiency.
Improved Documents, Slides and Design
Anthropic is also positioning Sonnet 5.5 as a practical model for knowledge work rather than focusing exclusively on programming.
The company says the model produces clearer writing and has improved at following visual templates and adding polish to user interfaces. Anthropic describes internal testing in which Sonnet 5.5 was given financial materials, earnings-call transcripts and a presentation template and asked to create a 10-slide operating review. Two experts judged the first draft ready to send without additional editing.
The company also reports improvements in long-horizon knowledge work, computer use and visual chart recognition.
On GDPval-AA v2.1, a benchmark covering real-world tasks across multiple occupations, Sonnet 5.5 scored 1,844 compared with 1,449 for Sonnet 5, while Opus 5.5 scored 1,846. Anthropic says the new Sonnet model therefore comes close to Opus 5.5 on this particular evaluation, although it continues to position Opus for more demanding open-ended work.
New Cybersecurity Safeguards
Claude Sonnet 5.5 also introduces a notable safety change within the Sonnet line.
Anthropic says the model’s cybersecurity capabilities have improved substantially compared with Sonnet 5. As a result, Sonnet 5.5 is the first Sonnet model to launch with cybersecurity safeguards and fallbacks similar to those used with Anthropic’s more capable models.
Routine software development remains supported, while higher-risk cybersecurity requests can fall back to Sonnet 5 under Anthropic’s safeguards.
Anthropic’s automated behavioral audit covered roughly 1,850 scenarios. The company reports that Sonnet 5.5 improved on or matched Sonnet 5 across most of its measures for alignment, resistance to misuse and honesty. Anthropic also cautions that evaluations cannot identify every possible model failure.
The model uses the same biology safeguards as Sonnet 5. Anthropic says these are designed to target harmful requests while leaving most research, education and clinical work unaffected.
Claude Sonnet 5.5 Availability
Claude Sonnet 5.5 is available immediately across Anthropic’s platforms, including Claude’s web and mobile applications and the Claude Platform for developers. It is also available through Amazon Web Services, Google Cloud and Microsoft Azure. Developers can access the model through the claude-sonnet-5-5 model identifier.
Anthropic says Sonnet 5.5 is also available with zero data retention, consistent with its previous Sonnet and Opus offerings.
The company has separately confirmed that Claude Haiku 5.5 is planned to join the Claude 5.5 family in the coming weeks. It is intended for high-volume and cost-sensitive applications, but it is not part of the current Sonnet 5.5 launch.
What Claude Sonnet 5.5 Means for Developers?
For developers, the main change with Claude Sonnet 5.5 is the combination of model performance and execution efficiency rather than a reduction in headline API token prices.
The same $2-per-million-input and $10-per-million-output pricing as Sonnet 5 remains in place, but Anthropic’s tests indicate that the newer model can complete comparable tasks using fewer tokens and produce results more quickly. That could make a difference for applications that run large numbers of coding, support, document-processing or AI-agent tasks.
The model also gives developers another option between routine, high-volume workloads and Anthropic’s more expensive Opus tier. Anthropic’s published results show Sonnet 5.5 approaching Opus 5.5 on several evaluations while maintaining lower token pricing, although the company continues to distinguish Opus 5.5 for complex, open-ended tasks requiring sustained judgment.
Claude Sonnet 5.5 is therefore primarily an efficiency-focused upgrade to Anthropic’s Sonnet line: faster generation, stronger coding and knowledge-work performance, unchanged headline token pricing, and new cybersecurity safeguards. The next major addition to the Claude 5.5 family will be Haiku 5.5, which Anthropic says is coming in the following weeks.
Friquently Asekd Questions (FAQs)
1. What is Claude Sonnet 5.5?
Claude Sonnet 5.5 is Anthropic’s second Claude 5.5 model. It is designed for everyday tasks, coding, documents, presentations, spreadsheets and other workloads where speed and efficiency are important.
2. How much does Claude Sonnet 5.5 cost?
The API price is $2 per million input tokens and $10 per million output tokens. Cache reads cost $0.20 per million tokens. Anthropic says the model can cost up to 30% less per task than Sonnet 5 because it generally uses fewer tokens.
3. Is Claude Sonnet 5.5 faster than Sonnet 5?
Yes. Anthropic reports that Sonnet 5.5 generates output more than 30% faster than Sonnet 5.
4. Is Claude Sonnet 5.5 available now?
Yes. Anthropic says Sonnet 5.5 is currently available across Claude and its developer platforms, including Amazon Web Services, Google Cloud and Microsoft Azure.
5. Is Claude Haiku 5.5 available?
Not yet according to the announcement. Anthropic says Claude Haiku 5.5 will join the Claude 5.5 family in the coming weeks.
Also Read –
Claude Opus 5.5: Everything You Need to Know About Anthropic’s AI Model
Claude Opus 5.5 Launches With Stronger Coding and Lower Costs
Source
Anthropic – Introducing Claude Sonnet 5.5
Anthropic – Claude Sonnet models
