Observed arrival · 2026-09-20
Apex Nexus Core: Small Models, Cast for One Job
An alpha-stage platform that turns a task description and examples into a small specialist language model you can own and run on your own hardware.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams running repetitive, data-sensitive language tasks
- Worth noticing
- The proposed cores target 1–8 billion parameters and ship in safetensors, GGUF, and ONNX formats.
Field notes
The proposed workflow begins by freezing a held-out test set, then uses large open-weight teacher models to synthesize task-specific data before training smaller candidates. Nexus says it will retain the smallest model that clears the target and evaluate it against regression and adversarial tests. The homepage lists safetensors, GGUF, and ONNX outputs, plus quantized builds, but presents these as pre-launch design targets.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue