Anthropic Blocks xAI from Claude; DeepSeek V4 Targets Code Supremacy at New Year

Anthropic announced a new policy that blocks all major competitors from using its Claude model, and this week it cut off access for Elon Musk’s xAI team, as revealed in an internal Slack message shared by xAI co‑founder Tony Wu.

“Hi team, many of you have probably noticed that Anthropic’s model no longer responds in Cursor. According to Cursor, this is a new policy Anthropic is applying to all major competitors.”

The shutdown follows previous actions by Anthropic, such as revoking OpenAI’s access before the GPT‑5 launch and restricting the Windsurf editor, indicating a pattern of “cleaning the house.”

Wu described the impact as both a “bad news” and a “good news” item: the immediate loss of productivity because Claude remains strong in programming tasks, but the forced restriction compels xAI to speed up development of its own coding models and products, a move he said will make the next year “very exciting.”

Bad news: short‑term productivity dip, acknowledging Claude’s current advantage in code generation.

Good news: pressure to accelerate xAI’s own coding AI efforts.

Musk also hinted on X that a major upgrade to Grok Code is coming next month, promising to “handle many complex coding tasks in one go.”

Meanwhile, Chinese AI startup DeepSeek is preparing to release its flagship model DeepSeek V4 in the next month, timed with the Chinese New Year. According to The Information, internal tests show V4’s code‑generation ability is “extraordinary” and even surpasses the current models from OpenAI and Anthropic.

A Twitter user @Yuchenj_UW commented that Anthropic’s restriction may accelerate other labs, noting the rumor that DeepSeek V4 could outperform Claude and GPT in coding.

The combined developments suggest a sharpening competitive landscape in AI code assistants, with established players tightening access while emerging Chinese models aim to leapfrog existing benchmarks.