Anthropic has completed training a new model — a potential successor to Mythos 5
An intriguing situation is brewing in the world of artificial intelligence. According to analytical community data, the company Anthropic has completed training on a new model that could become the direct successor to the Mythos family. While the corporation remains silent, neither confirming the existence of the new system, its name, nor its claimed specifications, insiders are already making predictions about its capabilities and market prospects.
Expert Andrew Curran, known for his accurate forecasts in the AI field, notes that the new system surpasses Mythos 5 in computing power and functionality. However, as he emphasizes, the final decision on the release remains with Anthropic. The model could be released under the name Mythos 5.1 or Mythos 6, or it could be kept entirely for internal use — within the framework of the Glasswing project, the company's internal cybersecurity platform.
Let me remind you that just nine days ago, on June 12, the U.S. Department of Commerce sent an order to Anthropic's management to suspend operations of the Mythos 5 and Fable 5 models. The reason was national security concerns. The restrictions affected all foreign nationals, including company employees born outside the United States. Anthropic was forced to disable both systems worldwide.
Fable 5, intended for public access, was equipped with multi-layered protection. Mythos 5, on the other hand, operated with fewer restrictions and was accessible through Glasswing. According to the company itself, using early versions of Mythos, about 50 partners have already identified over 10,000 critical and high-risk vulnerabilities in software.
Why the pause hasn't stopped progress
Curran believes that blocking development is unlikely to slow progress and could even accelerate it — by freeing up resources. As an example, he cites open-source competitors such as GLM-5.2 from Z.ai, which handles programming no worse than much more expensive closed models.
Thus, the fate of the next version remains undetermined. Whether Anthropic releases it publicly, restricts access to the Glasswing project, or keeps it only for internal use — this will significantly change the balance of power in the AI market. The company continues to seek restoration of access to both blocked models.
Expert opinion: The situation surrounding Anthropic is a vivid example of how regulatory pressure can both hinder and stimulate innovation. If the new model truly surpasses Mythos 5, we will witness a new round of the arms race in AI security. Investors and developers should closely monitor developments: the blocking of one model could open the door for even more powerful alternatives.