Disclaimer: This website is currently in beta testing. Some features may not be complete, and content may be out of date.
Back to Development updates
Roadmap·In development·5 min read

What's next: Ophraxx Lite • OX-B12 and OX-U45

Ophraxx Lite • OX-B12 and OX-U45 are the next two model tiers coming to the platform. Here's what each one is designed for, how the prompt classifier already prepares for them, and what the upgrade path looks like for server admins.

Abstract technology visualization

Ophraxx Lite • OX-B12: the reasoning model

Ophraxx Lite • OX-B12 is the next model in the lineup and the most immediate development priority. Where Lite OX-L1 is optimized for speed and brevity, Lite OX-B12 is built for depth — coding problems, multi-step analysis, technical explanations, essay drafts, and queries where accuracy matters more than response time. It targets a higher token output ceiling to support longer, more structured answers.

The prompt classifier we shipped in beta already knows about Lite OX-B12. Queries that score high on complexity — long messages, technical keywords, multiple questions — are already being routed toward the reasoning tier. Right now they fall back to Lite OX-L1 because Lite OX-B12 is not yet live. When it does launch, those queries will automatically go to the right model with no change required on the server admin side.

Ophraxx Lite • OX-U45: the flagship tier

Ophraxx Lite • OX-U45 is further out on the roadmap and represents the ceiling of what we plan to offer under the Ophraxx Lite name. It is designed for the most demanding long-form tasks — extended writing, complex multi-part reasoning, deep technical documentation. The timeline for Lite OX-U45 has not been set; Lite OX-B12 comes first.

All three tiers share the same safety stack, personality system, and conversation memory layer. The underlying model details are abstracted behind the Ophraxx Lite branding intentionally — we can upgrade what powers each tier without any change on the server side, and without disrupting existing configurations or expectations.

What this means for server admins

No reconfiguration is needed as new tiers launch. The routing system handles model selection automatically based on query complexity. When Lite OX-B12 goes live, complex queries in your server will simply start using it. If you want finer control — for example, biasing all queries toward the reasoning model regardless of complexity — that is something we are exploring as a future server-level setting.

Premium tiers will require a premium plan. The free tier will always include access to Ophraxx Lite • OX-L1. Details on premium pricing will be shared when Lite OX-B12 is closer to launch.