Anthropic Limited, Asian AI Startup Offers Counter Model

Illustration of an AI model. AI startups from China and Japan are launching new models amid export restrictions on advanced AI models from the United States. (Illustration by AI.)

Jakarta - AI startups from Japan and China have started to offer counter models after the United States government restricted global access to Mythos and Fable, two AI models owned by Anthropic.

TechCrunch, quoted Sunday, June 28, reported that Chinese cybersecurity company 360 introduced Tulongfeng, an AI tool that is claimed to be able to compete with Mythos. The Anthropic model focuses on cybersecurity and is said to be so powerful that access is restricted to non-American users.

In the same week, Tokyo's Sakana AI launched Fugu. Its name is taken from the Japanese word for pufferfish. The company calls Fugu an AI frontier model, a high-performance model at the forefront of AI development, which is on par with Anthropic's Fable 5 and Mythos Preview.

Fugu is also designed for AI agents. This model can regulate access to other models through an API or application programming interface.

Sakana stated that the launch of Fugu coincided with the Anthropic export ban, but not because of the ban. On its website, Sakana promotes Fugu as a frontier capable model without the risk of export control.

"Sakana Fugu is something we've been building on since last year," said a Sakana AI spokesperson. He said the timing happened to coincide with a moment that made Fugu get more attention than expected.

Sakana was founded in 2023 by former Google researchers Ren Ito, Llion Jones, and David Ha. The company makes generative AI models that are more affordable, can work with small data, and are optimized for Japanese languages and cultures.

Fugu is aimed at Japanese businesses and government agencies that want to reduce risk from tightening export controls. However, Sakana has not said Asia will abandon the American AI model.

"US models remain important for Asia," said Sakana's spokesman.

David Ha, co-founder and CEO of Sakana, said Fugu is not just a way to seize the market when US competitors are stuck with export rules. According to him, the orchestration model is the next stage because it can regulate the use of many AI models at once.

"Orchestration models are the next frontier, beyond the larger models," Ha wrote on X.

In China, 360 reportedly introduced two AI security tools. Tulongfeng is designed to automatically find software security gaps. Yitianzhen is made to automate cyber defense and incident response.

According to Reuters, quoted by TechCrunch, 360 founder Zhou Hongyi called AI security vulnerability searchers a national strategic asset. He also highlighted the risk of "one-way transparency", which is a condition when some parties have access to advanced vulnerability detection capabilities, while others do not.

Anthropic previously recorded significant growth. The US-based AI company said its annual revenue exceeded US$47 billion in May 2026. However, it is not known how large the company's customer base in Asia is.

Since the export restrictions came into effect, companies in Tokyo and Beijing have begun to offer alternatives to markets that previously relied on the Anthropic model. TechCrunch called local alternatives that are trained to understand local language and nuances are beginning to fill the space.