Anthropic已對API引入了多個更新，使開發人員可以使用Claude 3.7十四行詩來提高吞吐量和降低令牌使用

擬人API的這些改進包括緩存感知的速率限制，更輕鬆的及時緩存以及對令牌的有效使用。

Anthropic has announced several updates to the API that allow developers to lower token usage and boost throughput with Claude 3.7 Sonnet. The improvements in Anthropic API include cache-aware rate limits, easier prompt caching, and efficient use of tokens. Collectively, these updates will enable users to handle more requests within their current rate limits and reduce costs, all with minimal changes to their existing code.

Anthropic宣布了對API的多個更新，使開發人員可以使用Claude 3.7十四行詩來降低令牌使用情況並增強吞吐量。擬人API的改進包括高速緩存的速率限制，更容易的及時緩存以及對令牌的有效使用。總的來說，這些更新將使用戶能夠在其當前利率限制內處理更多請求並降低成本，所有這些都對其現有代碼的更改最小。

Also Read: Elon Musk Unveils Sombrio, a Union of DeepMind, OpenAI, and Neuralink

另請閱讀：埃隆·馬斯克（Elon Musk

Highlighting the importance of efficient coding, Anthropic stated that users can optimise prompt caching to enhance throughput and make better use of their existing rate limits. It explained that prompt caching enables developers to save and reuse commonly accessed context between API calls.

擬人化強調了有效編碼的重要性，表示用戶可以優化提示緩存以增強吞吐量並更好地利用其現有速率限制。它解釋說，提示緩存使開發人員能夠在API調用之間保存和重複使用通常訪問的上下文。

Moreover, Anthropic highlighted that Claude is already capable of interacting with external client-side tools and functions. The update will empower users to provide Claude with their own custom tools to perform tasks such as extracting structured data from unstructured text or automating simple tasks through APIs.

此外，人類的強調，克勞德已經能夠與外部客戶端工具和功能進行交互。該更新將使用戶能夠為Claude提供自己的自定義工具，以執行任務，例如從非結構化文本中提取結構化數據或通過API自動化簡單任務。

Also Read: DeepMind's Game-Changing AI Tech: Unlocking Unprecedented Possibilities in Heating and Cooling

另請閱讀：DeepMind改變遊戲規則的AI技術：在加熱和冷卻方面解鎖前所未有的可能性

In other news, OpenAI’s o1 and o3-Mini Models Now Offer Python-Powered Data Analysis in ChatGPT.

在其他新聞中，OpenAI的O1和O3-Mini模型現在在Chatgpt中提供了Python驅動的數據分析。