Trajectory launches post-training platform: supply tasks, data and eval criteria, deploy trillion-parameter models
udmrzn · x · 2026-11-04
Trajectory has launched a developer platform for post-training models, framing itself as the foundation for continual learning. Users provide the tasks the model should solve, the data needed, and evaluation criteria; the platform handles post-training, evaluation, and deployment — reportedly up to trillion-parameter models via a few agent prompts. Two research previews ship alongside: Continual grading, which scales a user's taste beyond what could be hand-labeled, and GS1, a new algorithm. The team envisions models that keep learning from the world after deployment.
Related event: Trajectory launches post-training platform for trillion-parameter models(4 posts)→
More from Models
- LLMs struggle with "surgicality": legal redlines leave unnecessary ink on the page — graceisford · 2026-11-04
- Frontier models race sailboats in a sim: Argon wins, Opus DNFs, Qwen runs aground — shekitup · 2026-11-04
- Domingos: Gemini itself agrees CoT and RL are a wasteful detour for reasoning — pmddomingos · 2026-11-04
- Legora CTO on frontier models that can understand and modify their own weights — garrytan · 2026-11-04
- LoL architecture reorders its own forward pass at runtime, mimicking algorithmic composition — micahgoldblum · 2026-11-04
- Grok Bot weekly quota 30% used within 9 hours of reset, user asks for bigger pool — nima_owji · 2026-11-04