Google announced Gemini 3.8 Flash Cyber and made it available through the Fairwind Program to 650 global partners including CrowdStrike, Datadog, Menlo Security, Palo Alto Networks, and Snowflake. Anthropic launched two new models: Claude Fable 5.1 and Claude Mythos 5.1, with Mythos available only through trusted access programs. Both companies offer different levels of safeguards and restrict access based on model sensitivity.
What’s Actually Happening
Google’s Fairwind program grants early access to advanced models for high-priority defence actors such as governments, healthcare providers, and telecommunications companies. This is not an open release for everyone. The Gemini 3.8 Flash Cyber model is designed specifically for detecting and responding to cyber threats, and is available only to verified organisations.
At Anthropic, Claude Fable 5.1 and Claude Mythos 5.1 offer different levels of capabilities and safeguards. Mythos, the more advanced model, is available only through trusted access programs - a mechanism that limits distribution to organisations meeting strict security and regulatory standards. This means not every organisation can simply purchase access.
Why It Matters Beyond the Headline
The shift to tiered access programs changes the rules for building enterprise AI systems. Previously, any organisation could purchase access to any public model. Now, organisations in critical sectors get a clear technological advantage - they see new capabilities and potential threats before everyone else.
In Israel, this is particularly relevant for organisations in digital health, telecommunications infrastructure, and government bodies. If you’re building an AI system that handles medical information or critical infrastructure, this early access could be the difference between detecting a threat in time and suffering a security incident. But it comes with a price: verification processes, detailed governance documentation, and sometimes external security audits.
Organisations not in these categories need to understand they’ll get access to these models later, if at all. This isn’t necessarily bad - models that have been field-tested by hundreds of large organisations arrive more stable and better documented.
What to Do This Week
If you’re in healthcare, telecom, or government, check whether you meet the criteria for the access programs. It’s not just a question of wanting in - you need appropriate security infrastructure, documented governance processes, and sometimes regulatory certifications. If you’re working with an external vendor to build AI systems, ask them whether they have access to these programs and what the integration process looks like.
If you’re not in these categories, don’t wait for these models. Plan your systems around the public models available today, and build them in a way that allows model replacement in the future. This means modular architecture, clear interfaces, and tests that don’t depend on a specific model. When the new models become available, you’ll be able to upgrade without rebuilding. For more on building such systems, see AI agents for business.
Sources
Frequently asked questions
Can a small business in Israel get access to Google's Fairwind program?
The program is designed for high-priority defence actors such as governments, healthcare providers, and telecommunications companies. A small business not in these categories probably won't get direct access, but could work with a technology provider that is a program member.
What's the difference between Claude Fable 5.1 and Claude Mythos 5.1?
The two models offer different levels of capabilities and safeguards. Mythos is the more advanced model and is available only through trusted access programs for organisations meeting strict security standards, while Fable is more widely available.