Meta delays latest AI model, Behemoth, says WSJ
Make us a preferred source
Make us a preferred source on Google
The news: Meta is delaying the rollout of its flagship “Behemoth” AI model due to concerns about its capabilities, The Wall Street Journal reported, citing unnamed sources familiar with the matter.
Engineers are struggling to significantly improve the capabilities of the large-language model, leading to internal questions about whether improvements over prior versions are significant enough to justify public release, according to the report. It added Behemoth’s performance has been hobbled by training challenges.
The context: Early in its development, Behemoth was internally slated for an April release to coincide with Meta’s inaugural AI conference for developers. Meta put out two smaller models in its Llama AI model family — Llama 4 Scout and Llama 4 Maverick — ahead of the event.
Behemoth was pushed to June, and is now delayed to autumn or later, according to the report.
The delay coincides with broader industry struggles, with OpenAI’s GPT-5 and Anthropic’s Claude 3.5 Opus also delayed. Senior executives are frustrated with the Llama 4 team and are contemplating significant management changes to the AI product group.
Meta has said Behemoth outperforms models from OpenAI, Google and Anthropic on some tests.
Meta has not publicly committed to a timeline but could release a more limited version sooner, the sources added. A Meta spokesman declined to comment to the journal.
The source: The Wall Street Journal
