doteval
doteval is an AI-assisted evaluation workspace that simplifies the creation of high-signal evaluations, alignment of LLM judges, and definition of rewards for reinforcement learning, all within a single platform. It offers a Cursor-like experience to edit evaluations-as-code against a YAML schema, enabling users to version evaluations across checkpoints, replace manual effort with AI-generated diffs, and compare evaluation runs on tight execution loops to align them with proprietary data. doteval supports the specification of fine-grained rubrics and aligned graders, facilitating rapid iteration and high-quality evaluation datasets. Users can confidently determine model upgrades or prompt improvements and export specifications for reinforcement learning training. It is designed to accelerate the evaluation and reward creation process by 10 to 100 times, making it a valuable tool for frontier AI teams benchmarking complex model tasks.
Learn more
Muse Spark 1.1
Muse Spark 1.1 is a multimodal reasoning model from Meta Superintelligence Labs built for agentic tasks, coding, computer use, tool use, and multimodal understanding. The model improves on the original Muse Spark with stronger performance in planning, orchestration, long-context work, coding workflows, and external app interactions. Muse Spark 1.1 can manage a 1 million token context window, remember earlier actions, retrieve important information, compact context, and delegate tasks across parallel subagents. It is designed to operate across tools, MCP servers, custom skills, browsers, native apps, scripts, images, video, PDFs, and audio-based workflows. Developers can access Muse Spark 1.1 through the new Meta Model API public preview, while users can try it in Thinking mode in the Meta AI app and on meta.ai.
Learn more
Peta
Peta is an enterprise-grade control plane for the Model Context Protocol (MCP) that centralizes, secures, governs, and monitors how AI clients and agents access external tools, data, and APIs. It combines a zero-trust MCP gateway, secure vault, managed runtime, policy engine, human-in-the-loop approvals, and full audit logging into a single platform so organizations can enforce fine-grained access control, hide raw credentials, and track every tool call made by AI systems. Peta Core acts as a secure vault and gateway that encrypts credentials, issues short-lived service tokens, validates identity and policies on each request, orchestrates MCP server lifecycle with lazy loading and auto-recovery, and injects credentials at runtime without exposing them to agents. The Peta Console lets teams define who or which agents can access specific MCP tools in specific environments, set approval requirements, manage tokens, and analyze usage and costs.
Learn more
Muse Spark 1.2
Muse Spark 1.2 is Meta’s coding-focused model update designed to power Muse Code and improve software engineering workflows. The model is built for code generation, complex debugging, codebase understanding, long-horizon development tasks, and end-to-end developer workflows. Muse Spark 1.2 was co-trained with Muse Code to improve performance inside the terminal coding agent environment. It supports planning, goal conditioning, context compaction, subagent coordination, and iterative coding workflows across large repositories. The model was trained with expanded coding compute, diverse development environments, self-improvement loops, and long-running engineering tasks. Built for AI developers and software teams, Muse Spark 1.2 helps agents plan, write, validate, debug, and optimize code with greater autonomy.
Learn more
