Google, OpenAI, Anthropic all launch new models – none fully available yet

Google’s Gemini 3.5 Pro, announced at I/O on May 19 with a 2-million-token context window and 'Deep Think' reasoning mode, has slipped from June to July for general availability. Only a limited enterprise preview is currently live. OpenAI’s GPT-5.6 generation is rolling out in a mainstream version, but its top-tier Sol variant—targeting hard coding and security tasks—is restricted to roughly 20 organizations after coordinating with the US government. Anthropic’s Claude Fable 5, a near-flagship model sitting below Mythos 5, was pulled offline in June under an export-control directive and is only returning on July 1. All three labs have avoided publishing model cards or public benchmarks, citing limited release windows.

Because none of these models are fully accessible or benchmarked independently, no clear winner can be declared. Previous generation trends suggest Google leads in breadth and long context, Anthropic in code quality, and OpenAI in math and consumer reach—but whether that holds for these new models remains unconfirmed. Enterprises should watch for wider releases in the coming weeks, but currently cannot rely on any single model as a general-purpose solution.