Anthropic introduces Claude Sonnet 5.5: Faster AI model promises lower costs
Anthropic’s latest Claude model focuses on faster responses, stronger coding capabilities and improved efficiency, while maintaining the same token pricing across its input, output and cache-read tiers.
- Sep 29, 2026,
- Updated Sep 29, 2026 9:13 AM IST
Anthropic has launched Claude Sonnet 5.5, an improved version of its Claude Sonnet 5 model that emphasises speed, efficiency, and better performance for coding and other knowledge-work tasks. This new model is intended for everyday work such as coding, creating documents, and completing complex tasks, while requiring fewer resources. According to Anthropic, Claude Sonnet 5.5 produces responses more than 30% faster than the previous version and can cost up to 30% less per task even though it has the same token pricing.
The company is presenting the model as a faster, less expensive alternative to Claude Opus 5.5, especially for well-defined tasks that don't require the continuous judgment characteristic of its top model. It is also the first Sonnet model to come with enhanced cybersecurity measures and fallback options.
Claude Sonnet 5.5 gets a major coding boost
Anthropic says the new model has made significant progress in coding. Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0, up from 10.3% for Sonnet 5, and achieved 55.5% on CursorBench 4.0, with improvements in both long-horizon knowledge work and image understanding. Anthropic adds that it is the first Sonnet model to beat Pokémon Red using only screenshots.
The company also points out improvements in the area of collaboration, stating that the model now writes more clearly and is more appropriate for rapid iterations. Early testers noted that it can understand codebases quickly and needs fewer steps and tool calls.
Same token price, but lower cost per task
For you, the main difference is efficiency, not price. Claude Sonnet 5.5 has a cost of $2 per million input tokens, $10 per million output tokens and $0.20 per million cache-read tokens - the same rates as Sonnet 5. Nevertheless, Anthropic says the model generally uses fewer tokens to do the same task, reducing cost per task by as much as 30%.
Sonnet 5.5 is available on all of Anthropic's platforms, such as Amazon Web Services, Google Cloud, and Microsoft Azure, and developers can use it via the Claude Platform using the model identifier claude-sonnet-5-5.
For Unparalleled coverage of India's Businesses and Economy – Subscribe to Business Today Magazine
Anthropic has launched Claude Sonnet 5.5, an improved version of its Claude Sonnet 5 model that emphasises speed, efficiency, and better performance for coding and other knowledge-work tasks. This new model is intended for everyday work such as coding, creating documents, and completing complex tasks, while requiring fewer resources. According to Anthropic, Claude Sonnet 5.5 produces responses more than 30% faster than the previous version and can cost up to 30% less per task even though it has the same token pricing.
The company is presenting the model as a faster, less expensive alternative to Claude Opus 5.5, especially for well-defined tasks that don't require the continuous judgment characteristic of its top model. It is also the first Sonnet model to come with enhanced cybersecurity measures and fallback options.
Claude Sonnet 5.5 gets a major coding boost
Anthropic says the new model has made significant progress in coding. Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0, up from 10.3% for Sonnet 5, and achieved 55.5% on CursorBench 4.0, with improvements in both long-horizon knowledge work and image understanding. Anthropic adds that it is the first Sonnet model to beat Pokémon Red using only screenshots.
The company also points out improvements in the area of collaboration, stating that the model now writes more clearly and is more appropriate for rapid iterations. Early testers noted that it can understand codebases quickly and needs fewer steps and tool calls.
Same token price, but lower cost per task
For you, the main difference is efficiency, not price. Claude Sonnet 5.5 has a cost of $2 per million input tokens, $10 per million output tokens and $0.20 per million cache-read tokens - the same rates as Sonnet 5. Nevertheless, Anthropic says the model generally uses fewer tokens to do the same task, reducing cost per task by as much as 30%.
Sonnet 5.5 is available on all of Anthropic's platforms, such as Amazon Web Services, Google Cloud, and Microsoft Azure, and developers can use it via the Claude Platform using the model identifier claude-sonnet-5-5.
For Unparalleled coverage of India's Businesses and Economy – Subscribe to Business Today Magazine