Anthropic Faces Backlash as the Only Major AI Lab Refusing to Support Open Models

Anthropic is facing industry criticism for its position on open models, with Business Insider reporting that it is now the only major AI lab not supporting open-source initiatives. As competitors pivot toward more open strategies, Anthropic’s closed-source approach is attracting growing attention and pushback.

The report notes that OpenAI and Google have both embraced open-source strategies to varying degrees in recent years — OpenAI has released several open models, and Google has opened up its Gemma series. In contrast, Anthropic has never released any open-source version of its Claude model, making its position stand out sharply in the industry.

Critics argue that in an era where AI safety discussions increasingly emphasize transparency and external auditability, fully closed-source approaches hinder independent researchers from conducting safety evaluations and red-teaming exercises. The open-source community has also voiced concerns that the lack of open models limits academic research and small-scale developers’ access to frontier AI capabilities.

Anthropic has previously emphasized its commitment to AI safety, arguing that controlled release and safety evaluation are preferable to open sourcing for managing model risks. The company has argued that open sourcing could introduce uncontrollable safety risks, especially as model capabilities continue to advance.

However, as OpenAI and Google push forward with open-source initiatives and Meta’s Llama series continues to evolve, Anthropic’s closed-source stance faces growing market pressure. Developer communities and enterprise users increasingly consider open-source availability as a key factor when choosing AI platforms.

The controversy reflects a fundamental tension in the AI industry: how to balance safety with openness. Whether Anthropic will adjust its strategy or find a new middle path between security and transparency will be closely watched by the industry.