Anthropic’s latest AI model is making a case that businesses no longer have to choose between performance and affordability.
The company unveiled Claude Sonnet 5 on June 30, describing it as its most agentic Sonnet model yet. According to Anthropic, the model can plan tasks, use tools such as browsers and terminals, and run autonomously at a level that previously required larger, more expensive models.
The launch reflects a broader shift in the AI industry. Instead of competing solely on benchmark scores, vendors are racing to build AI that businesses can actually afford to deploy at scale. Here’s what you need to know about Claude Sonnet 5 and what it could mean for enterprise AI.
What is Claude Sonnet 5?
Claude Sonnet 5 is Anthropic’s latest Sonnet model, designed to balance performance, speed, and cost.
Unlike larger frontier models that prioritize maximum capability, Sonnet is intended to offer strong performance at a lower price. Anthropic says Sonnet 5 narrows the gap with Claude Opus 4.8 while improving over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work.
Claude Sonnet 5 is available across all Claude plans. It is the default model for Free and Pro users and is available to Max, Team, and Enterprise users. Developers can also access it through Claude Code, the Claude Platform, and the Claude API.
What’s new in Claude Sonnet 5?
Anthropic says Claude Sonnet 5 introduces improvements across several areas:
- Stronger reasoning for complex business tasks
- Better coding performance and software development assistance
- Improved use of tools such as web browsers and terminals
- Enhanced support for long-running, multi-step AI agents
- Better instruction following and task planning
- Updated safety protections and improved resistance to prompt injection attacks
| Reasoning | Improved over Sonnet 4.6 | Helps businesses handle more advanced analysis, planning, and decision support |
| Coding | Stronger software development assistance | Could support debugging, code generation, testing, and developer productivity |
| AI agents | Better support for long-running, multi-step workflows | Makes AI automation more useful for business processes that require planning and tool use |
| Tool use | Improved browser, terminal, and external tool interactions | Allows the model to complete work across systems instead of only responding in chat |
| Pricing | Lower introductory API pricing through Aug. 31, 2026 | Makes large-scale AI pilots and agent deployments easier to justify financially |
| Safety | Better refusal behavior and prompt injection resistance | Reduces risk for businesses using AI in sensitive or regulated environments |
Together, these updates move Claude further beyond chatbot-style interactions and toward autonomous systems capable of completing real business workflows.
Why is everyone talking about AI agents?
One of the biggest themes of the announcement is agentic AI.
Traditional chatbots answer questions one prompt at a time. AI agents are designed to complete multi-step tasks with minimal supervision. They can plan work, gather information, use software tools, execute code, and adapt as conditions change.
Anthropic says Claude Sonnet 5 was designed specifically for these types of workflows, making it a stronger candidate for enterprise automation projects.
For businesses, that could mean AI systems capable of troubleshooting software, analyzing documents, researching information, generating code, or completing repetitive administrative tasks with fewer manual prompts.
How much does Claude Sonnet 5 cost?
Pricing may be one of Anthropic’s most important announcements.
Claude Sonnet 5 launches with introductory pricing of $2 per million input tokens and $10 per million output tokens through Aug. 31, 2026. After that, pricing increases to $3 per million input tokens and $15 per million output tokens.
There is one important caveat: Anthropic says Sonnet 5 uses an updated tokenizer, which changes how the model processes text. The same input can map to roughly 1.0 to 1.35 times more tokens, depending on the content type. Anthropic said the introductory pricing was designed to make the transition to Sonnet 5 roughly cost-neutral.
Those prices are especially important for organizations deploying AI agents, which typically consume far more tokens than traditional chatbot interactions because they plan, reason, verify results, and repeatedly interact with external tools.
Lower operating costs could make enterprise AI projects easier to justify financially.
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
How does Claude Sonnet 5 compare with Opus?
Anthropic positions Sonnet as the practical alternative to its flagship Opus models.
According to the company, Claude Sonnet 5 performs close to Claude Opus 4.8 while costing less. Opus 4.8 remains Anthropic’s higher-accuracy option for some agentic search and computer use tasks, but Sonnet 5 offers a lower-priced option that is much stronger than previous Sonnet models.
For many businesses, that raises an important question: Do they actually need a flagship model?
If Sonnet delivers sufficient performance for common business applications at a lower cost, it may become the more practical choice for enterprise deployments.
What about safety?
Anthropic continues to differentiate itself through its focus on AI safety.
The company says Claude Sonnet 5 showed a lower overall rate of undesirable behavior than Sonnet 4.6 during internal testing. It also reported improvements in rejecting malicious requests, resisting hijack attempts during prompt injection attacks, and reducing hallucinations and sycophancy.
Anthropic also says the model can perform some routine, non-harmful cybersecurity tasks, but it remains substantially less capable than current Opus models on dangerous cyber evaluations, such as developing software exploits. The company said Sonnet 5 was unable to develop a fully working exploit in a single Firefox vulnerability evaluation, although it showed a slightly higher partial success rate than Sonnet 4.6.
Because Sonnet 5 outperforms its predecessor on some cyber-related tasks, Anthropic launched it with cyber safeguards enabled by default.
For regulated industries and security-conscious organizations, those safeguards could influence purchasing decisions just as much as benchmark performance.
What does this mean for IT leaders?
Claude Sonnet 5 is another sign that enterprise AI is entering a new phase.
Organizations are increasingly evaluating AI models based on practical business outcomes rather than raw benchmark performance. Questions about deployment costs, security, governance, and integration are becoming just as important as reasoning scores.
For IT leaders, now is a good time to revisit existing AI pilots and proof-of-concept projects. A model that combines stronger agentic performance with lower operating costs could make automation viable for software development, IT operations, customer support, knowledge management, and other business functions.
The more important question is no longer, “Which model is the smartest?” Instead, it’s “Which model delivers the most value for our business?”
As Anthropic, OpenAI, Google, Microsoft, and other AI vendors continue competing on both capability and cost, that answer may become increasingly favorable for enterprise customers.
Frequently asked questions
Is Claude Sonnet 5 available now?
Yes. Anthropic says Claude Sonnet 5 is available across all Claude plans. It is the default model for Free and Pro users and is available to Max, Team, and Enterprise users. It is also available in Claude Code, on the Claude Platform, and through the Claude API.
Is Claude Sonnet 5 free?
Claude Sonnet 5 is available to Free and Pro users as the default Claude model. Developers and businesses using the Claude API pay based on token usage.
How much does Claude Sonnet 5 cost?
Claude Sonnet 5 costs $2 per million input tokens and $10 per million output tokens through Aug. 31, 2026. After that, it will cost $3 per million input tokens and $15 per million output tokens.
Anthropic notes that Sonnet 5 uses an updated tokenizer, so the same input may map to more tokens than before.
Is Claude Sonnet 5 better than Opus?
Not across the board. Anthropic says Sonnet 5 performs close to Claude Opus 4.8 at a lower price, but Opus 4.8 remains the better choice for higher accuracy on some agentic search and computer use tasks.
Should businesses switch to Claude Sonnet 5?
Not necessarily. Organizations already invested in another AI platform should compare performance, pricing, security features, governance needs, and ecosystem integrations before making changes. However, businesses evaluating AI agents may want to include Claude Sonnet 5 in future testing.
Related: Want to know how Anthropic plans to power Claude’s next generation? Read our coverage of its partnership with SpaceX to tap the massive Colossus 1 AI supercomputer.