Crypto Firms Urge AI Labs to Share Frontier Models with Defenders
A coalition of crypto companies and the Bitcoin Policy Institute have called on major AI labs to grant trusted-access programs for qualified defenders, giving them early access to frontier models. This comes as cyber threats become more advanced, with open-source maintainers often lacking access to advanced cyber programs run by AI labs.
The group argues that stronger tools could help defenders identify vulnerabilities before attackers exploit open-source financial infrastructure. Bitcoin alone secures over $1 trillion in value, and the coalition wants smaller nonprofits, independent maintainers, and open-source security teams to qualify for advanced defensive systems.
OpenAI has already expanded its Daybreak program with Daybreak Blue and Daybreak Red for approved cybersecurity defenders. Its GPT-5.6-Cyber model completed 95% of requests in an internal advanced cybersecurity evaluation, while the standard GPT-5.6 Sol model only managed 1.5%.
The industry has also increased funding for Bitcoin security, with a $15 million effort to support Bitcoin developers and security researchers launched by Strategy, BlackRock, and Coinbase last July. The coalition's request targets similar defenders seeking access to stronger cybersecurity tools, not unrestricted access to offensive-capable AI models.