Zhou Hongyi, founder of Chinese cybersecurity giant Qihoo 360, made a pointed claim this week: China now has its own equivalent to Anthropic's Mythos, the advanced AI model the U.S. government locked behind export controls on June 12. One version of China's homegrown alternative, he says, is free.
The timing is not accidental.
Since the U.S. restricted Anthropic's Fable 5 and Mythos 5 from foreign nationals following a disputed jailbreak finding, the scramble to fill that gap has accelerated on the Chinese side. Mythos was specifically flagged for its cybersecurity capabilities, which is exactly the domain where Qihoo 360 has spent decades building expertise. Zhou's claim lands with that context fully intact.
What Qihoo 360 is actually claiming
Here's the thing: Zhou is not just making a vague nationalistic boast. The assertion is specific. China's version reaches capability parity with Mythos on the tasks that triggered U.S. concern in the first place, and the free tier makes it accessible in a way that Anthropic's restricted models simply cannot be for international users right now.
Qihoo 360 built its reputation on antivirus software and network security, which gives the company a direct line to the exact use cases Mythos was designed for. The free version of their model is positioned as a direct answer to the access vacuum the export order created.
What most players miss here is the strategic framing. This is not just a product launch. It is a public signal that U.S. export controls, intended to slow the spread of frontier AI capabilities, may be accelerating domestic development in China instead. The restriction created demand. Qihoo 360 is claiming to supply it.
Why this matters beyond the AI industry
The broader context here connects directly to what Anthropic is navigating with its own model lineup. Claude Sonnet 5 launched on June 30 specifically without the cybersecurity training that got Mythos flagged, and it ships without the export restrictions. That is not a coincidence. Anthropic is threading a needle between capability and compliance.
Meanwhile, Chinese labs are not operating under the same constraints. Zhou's announcement is the clearest public example yet of how that asymmetry plays out. When a frontier capability gets restricted in one market, the incentive to build a domestic version in another market does not disappear. It intensifies.
The free tier is the key here. Paid access creates friction. Free access creates adoption, and adoption at scale is how a model's capabilities get tested, refined, and eventually surpassed. If Qihoo 360's model genuinely operates at Mythos-level on cybersecurity tasks and is available without cost, the downstream effects on the competitive gap between U.S. and Chinese AI development become significantly harder to measure.
The free model play is a familiar move
The strategy of releasing a capable model for free to drive adoption is not new. It is how Chinese labs have consistently challenged the assumption that frontier AI requires frontier pricing. DeepSeek ran the same playbook earlier this year, and the reaction from U.S. labs was immediate and visible.
Qihoo 360 entering that conversation with a Mythos-level claim and a free tier is a direct continuation of that pattern. The difference this time is the explicit positioning against a model that is actively restricted by government order. That framing gives the announcement a dimension that a standard product launch would not carry.
For gamers and developers tracking AI tools for game development, coding assistance, and agentic workflows, the practical upshot is that the pool of capable free-tier models is growing. If you want to stay current on what these tools can actually do in practice, the Mythic Protocol starter guide is a solid example of how AI-assisted documentation is already changing how players approach complex systems. For a broader look at what the current generation of AI-adjacent gaming tools supports, the gaming guides hub tracks the latest across genres.
The next few months will show whether Zhou's claim holds up under scrutiny. Benchmark results from independent evaluators, not the company itself, will be the real test. Until then, the statement alone has already done its work: China has named its Mythos, and it is free to use.
