Anthropic updates Sonnet 5.5: prices are unchanged, but completing the same work costs less Token

Article

Anthropic updates Sonnet 5.5: prices are unchanged, but completing the same work costs less Token Beating news flash:AI Anthropic has released Claude Sonnet 5.5, continuing to focus on everyday work and programming.API Unit pricing remains unchanged from Sonnet 5, with input and output priced per million tokens at 2 USD and 10 USD, exactly half the price of Opus 5.5. Anthropic says the new model generates output 30%faster and typically requires fewer tokens to complete the same task, reducing per-task costs by up to 30%.

Agentic programming saw the most significant improvement.Terminal-Bench 4.0(where the model completes multistep tasks in the command line), Sonnet 5.5 rose from Sonnet 5's 10.3%to 70.6%, surpassing Opus 5.5's 66.4%. However, among the 8 evaluations published by the company, Sonnet 5.5 outperformed Opus only in this category; Opus 7 retained an advantage in the other 5.5 categories, although most gaps are now small. Anthropic also explicitly stated that complex, open-ended tasks requiring sustained judgment remain better suited to Opus 5.5.

Yet, on a per-task basis, costs may not necessarily be cut in half. At the highest Max reasoning setting, Artificial Analysis found that Sonnet 5.5's Intelligence Index was 56，Opus 5.5 Max was 58. However, Sonnet 5.5 used an average of approximately 193000 output tokens per task, bringing its per-task cost in this test suite to 7.60 USD, which was instead higher than Opus 5.5 Max at 5.98 USD. Whether it saves money in practice depends heavily on the task and reasoning setting.

Sonnet 5.5 is now available on Claude, Claude API、AWS, Google Cloud, and Microsoft Foundry. Haiku 5.5 will be released in the coming weeks.

Original link https://m.theblockbeats.info/flash/369458

Key points

Anthropic releases Claude Sonnet 5.5, continuing to focus on everyday work and programming.
API Input pricing is 2 USD per million tokens, while output pricing is 10 USD per million tokens, unchanged from Sonnet 5.
Anthropic says the new model generates output 30%faster, with per-task costs reduced by up to 30%。
Terminal-Bench 4.0, Sonnet 5.5 scored 70.6%, above Sonnet 5's 10.3%and Opus 5.5's 66.4%。
Among the 8 evaluations published by the company, Sonnet 5.5 led only in Terminal-Bench 4.0, ahead of Opus 5.5。
Artificial Analysis found that Sonnet 5.5 used an average of approximately 193000 output tokens per task, at a per-task cost of 7.60 USD。
Sonnet 5.5 is now available on Claude, Claude API、AWS, Google Cloud, and Microsoft Foundry.
Haiku 5.5 will be released in the coming weeks.

AI-assisted interpretation

The following is analysis, separate from reported facts. Verify important claims independently.

This report states that Sonnet 5.5's listed price has not decreased, but it may use fewer Token when completing tasks, potentially lowering total costs for some tasks. Test results show strong performance on multistep command-line tasks; however, complex, open-ended tasks remain better suited to Opus 5.5, and actual costs will also vary by task and reasoning setting.

Why it matters to readers

Model selection should not be based solely on API unit pricing; it should also consider the Token volume required to complete the same task, the task type, and the reasoning setting.

If you use Claude or related API，Sonnet 5.5, it may be faster or more cost-efficient for certain programming tasks, but this does not mean that every task will be less expensive.

Risks and unknowns

The original article does not provide a link to an official Anthropic announcement or pricing page; the version, pricing, and availability scope still require verification.
The per-task cost data comes from Artificial Analysis testing at a specific Max reasoning setting and is not representative of all tasks.
The original article does not specify the particular tasks and test conditions to which “reduced by up to 30%” applies.
Haiku 5.5's specific release date and pricing have not yet been provided.

Anthropic updates Sonnet 5.5: prices are unchanged, but completing the same work costs less Token

Article

AI-assisted interpretation

Why it matters to readers

Risks and unknowns

Related Developments

Related concepts

Keep following

Security Center