Prime Intellect Releases prime-rl 0.6.0 to Train Trillion-Parameter MoE Models on Agentic RL Workloads
Prime Intellect has released prime-rl version 0.6.0. The framework targets reinforcement learning on trillion-parameter Mixture-of-Experts (MoE) models. It focuses on heavy agentic workloads, like long-horizon software-engineering tasks. The research team trained GLM-5 on SWE tasks at up to 131k sequence length. Step times stayed under five minutes. The batch size was 256 rollouts. The run used only 28 H200 nodes. TL;DR prime-rl 0.6.0 trains trillion-parameter MoE models on agentic RL
Get smart on it
Prime Intellect released an open framework called prime-rl 0.6.0 that trains very large AI models on reinforcement learning tasks, specifically focusing on long-horizon software engineering work. The framework separates the training and inference systems to run independently, avoiding delays when some tasks take much longer than others to complete. Key technical innovations include using lower-precision math (FP8) for faster processing, spreading experts across many GPUs, separating prefill and decode operations, and managing large memory caches through tiering. The approach demonstrated the ability to train a trillion-parameter model on complex coding tasks with reasonable resource requirements.
Related news
Mistral OCR 4 Brings Citation-Ready Structured Output to RAG, Agentic, and Enterprise Search Pipelines
Today, Mistral AI released OCR 4, its latest document-understanding model. This new release adds bounding boxes, block classification, and inline confidence scores alongside extracted text. It supports 170 languages across 10 language groups and runs in a single container for fully self-hosted deployments. OCR 4 also serves as an ingestion component for enterprise search, RAG, and domain-specific retrieval pipelines. TL;DR OCR 4 returns bounding boxes, typed-block labels, and per-word c
Datalab Releases lift: A 9B Open-Weights Vision Model That Extracts Structured JSON From PDFs Using Schemas
Datalab has released lift, a 9B open-weights vision model for structured extraction. You pass it a JSON schema, and it returns a JSON object that matches. The model reads PDFs and images directly, then decodes against your schema. This is Datalab’s first model built purely for extraction. The team already ships open-source OCR tools: chandra, marker, and surya. lift extends that work into schema-driven field extraction. lift scores 90.2% field accuracy on Datalab’s 225-documen
