TokenAI teases Horus 1.5, 64K context

This reads less like a launch and more like TokenAI signaling it wants Horus to become a real model family, not just a one-off regional demo.

• –The official Horus 1.0 materials describe a from-scratch, open-source model, which gives the project more credibility than a typical “local AI model” announcement.
• –A jump from 8K to 64K context is the kind of upgrade that matters for developer workflows, retrieval, and longer agentic tasks.
• –The cybersecurity-model tease is strategically interesting: security-focused LLMs can be more defensible than generic chat models if the data and evals are real.
• –The biggest question is execution, not ambition; the post makes strong claims on quality and scale, but there’s no independent validation yet.
• –If TokenAI can keep shipping in public, Horus could become a notable Arabic-region open-model effort rather than just a national first.