SWE-2
SWE-2 is Cognition’s advanced coding model designed to improve software engineering performance while reducing the cost of agentic coding workflows. The model is post-trained from Kimi K3 and uses reinforcement learning to optimize multiple reasoning-effort levels within a single training run. SWE-2 is designed to explore codebases more selectively, begin implementation sooner, and complete tasks with fewer redundant reads and reasoning steps than earlier Cognition models. Its capabilities include code generation, debugging, test creation, verification, repository analysis, and complex terminal-based software engineering tasks. The model also emphasizes stronger engineering judgment, end-to-end test coverage, instruction following, and evidence-based verification of user assumptions. SWE-2 is available through Devin Desktop and Devin CLI, with broader rollout planned across Devin Web and Fusion.
Learn more
Prime Intellect
Prime Intellect is the open superintelligence stack: an integrated compute, training, inference, and sandbox platform for teams that want to train, deploy, and continuously improve their own models. The stack is built around owning intelligence instead of waiting on frontier models to improve, giving users one loop for reinforcement learning environments, hosted evaluations, large-scale training, inference, and compute. In Lab, teams can post-train self-improving agents by turning tasks into RL environments, creating, developing, evaluating, and pushing them with the Prime CLI. The Environment Hub gives access to and contributions across 2,500+ open-source RL environments, while hosted evaluations let teams benchmark model performance across open-source models with no infrastructure or setup. Hosted Training supports large-scale models optimized for agentic workflows, managed training workflows with full visibility and control, and hands-on support from the applied research team.
Learn more
OORT DataHub
Data Collection and Labeling for AI Innovation.
Transform your AI development with our decentralized platform that connects you to worldwide data contributors. We combine global crowdsourcing with blockchain verification to deliver diverse, traceable datasets.
Global Network: Ensure AI models are trained on data that reflects diverse perspectives, reducing bias, and enhancing inclusivity.
Distributed and Transparent: Every piece of data is timestamped for provenance stored securely stored in the OORT cloud , and verified for integrity, creating a trustless ecosystem.
Ethical and Responsible AI Development: Ensure contributors retain autonomy with data ownership while making their data available for AI innovation in a transparent, fair, and secure environment
Quality Assured: Human verification ensures data meets rigorous standards
Access diverse data at scale. Verify data integrity. Get human-validated datasets for AI. Reduce costs while maintaining quality. Scale globally.
Learn more
Perle
Perle is a Web3-powered AI data platform designed to improve how artificial intelligence models are trained by combining human expertise with blockchain-based verification and incentives. It enables contributors to review, label, and evaluate multimodal data such as text, images, video, audio, and code, transforming human knowledge into structured, high-quality datasets used in real AI systems. It connects enterprises and AI labs with a global network of qualified contributors, ensuring that data used for training is accurate, context-rich, and aligned with domain expertise. Perle emphasizes data quality through multi-layer validation pipelines and consensus mechanisms that elevate annotation accuracy to production standards. Every contribution is recorded on-chain using the Solana blockchain, creating an immutable and transparent record of who contributed, what was done, and how it was validated, which improves trust, auditability, and compliance.
Learn more
