Meta Platforms has delayed the release of its flagship artificial intelligence model, «Behemoth,» citing concerns over its capabilities. Originally slated for an April launch, the release was first postponed to June and is now expected to be delayed until at least the fall, according to sources familiar with the matter.
The decision to postpone comes amid internal assessments indicating that Behemoth does not yet offer significant advancements over existing models. This has led to questions about the scalability of current AI technologies and whether the industry is approaching the limits of performance improvements through scaling alone.
Despite the delay, the impact on businesses is expected to be minimal, as many already utilize Meta’s earlier open-source models, such as Llama 4. The postponement, however, underscores the challenges tech companies face in pushing the boundaries of AI capabilities and the importance of ensuring robust performance before public deployment.
