Born Applied Intelligence · release-led model factory
Born-9B Preview
is live.
Born is the functional AI lab behind LeemerLabs. Right now the story starts with a real release: Born-9B Preview, the evals behind it, the chat surface it feeds, and the next general release path already being shaped in public.
Current release position
Born-9B Preview v2
0.9244
local weighted score
23 / 25
fixed held-out pass count
0.8559
SWE proxy score
7,097
v2 prepared rows
Later v2-recovery, v2.2/v2.3 hotfix, and preview recovery runs are all kept visible, but none beat the promoted v2 checkpoint on the same gate.
Born Chat is already open on top of this model lane while the general-release work shifts toward narrower, executable, test-backed improvements instead of broad “more data” passes.
Public preview adapter: rk500/Born-9B-Qwen3.5-9B-Preview
Born-9B Preview
The promoted preview stays on the v2 adapter. Current release position: 0.9244 on the fixed local 25-task gate, 23/25 passing, with later recovery and hotfix runs preserved but not promoted.
Born Chat
PromptKit UI, live model lanes, login, quotas, uploads, and a public training loop notice. This is where people can actually use the current Born lanes instead of reading around them.
Research and eval
Research is now the merged surface for company context, method, benchmarking, and release logic, instead of scattering those pieces across several thin pages.
LeemerLabs access
LeemerLabs handles inference and API access. Born handles the model work behind it. The split is deliberate, and both sides now read clearly in the product hierarchy.
From the lab
New notes, release work,
and what is moving now.
Born-9B Preview is finally here
The launch note for Born-9B Preview: Repath at the controls, GPT-5.5 Codex managing the run, the datasets, failures, benchmarks, and the public adapter.
Born Chat is open: PromptKit UI, live model lanes, and the public training loop
What shipped in Born Chat, how the Born and comparison lanes are organized, why the training notice is explicit, and what Born Chat Pro is meant to become.
Why Codex GPT-5.5 Operated Born-9B
The operating model behind the run: Codex GPT-5.5 High, Repath Ray Khan at the controls, and a model release managed through evidence, caveats, and repository state.
How Born-9B Learned to Breathe
The origin story of Born-9B told from the real notes: the cheap first run, the honest tie, the many teachers, and the million-token second inhale.
What Born does
Not just one model.
A working applied stack.
The release page is the artifact. The research page is the method and evaluation frame. The services page is the commercial surface. Together they explain Born without forcing visitors through duplicate company pages.
Model adaptation
LoRA, QLoRA, continuation runs, and release packaging built around task fit rather than generic tuning claims.
Dataset engineering
Synthetic generation, teacher distillation, provenance, splits, and validation before a row ever reaches training.
Evaluation systems
Local held-out gates, proxy issue-resolution packs, BornBench lanes, and release rules that stop claims outrunning evidence.
Agent workflows
Tool-use closure, verification loops, repo-aware repair, and the operational glue around the model itself.
Use the current system
Chat against the live lanes,
then read how they were built.
Born now reads as a smaller, cleaner system: one release-led home, one deep research page, one flagship model dossier, and one services page. The rest should support those surfaces, not compete with them.