On Tuesday, Anthropic cut off OpenAI’s access to its API for Claude models, citing violations of usage terms. Anthropic spokesperson Christopher Nulty noted that OpenAI staff improperly used Claude Code, a preferred tool for coders, ahead of GPT-5’s launch. The terms prohibit creating alternative products or replicating Anthropic’s technology. Insiders revealed OpenAI accessed Claude through custom APIs, allowing comparisons with its own models, raising safety concerns. OpenAI’s Chief Communications Officer, Hannah Wong, expressed disappointment over the API restriction, reinforcing that benchmarking is standard practice in the industry. Following this, Anthropic plans to impose weekly usage caps on Claude Code starting August 28, affecting all subscription tiers. These limits aim to manage excessive consumption, which led to outages due to high demand. Despite the service’s popularity, Anthropic continues to face challenges in scalability and resource allocation.
Source link
Share
Read more